Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
APPROVED:
Robert C. Benjamin, Major Professor Laura Gahn, Committee Member Joseph Warren, Committee Member Harrell Gill-King, Committee Member Jeff Johnson, Committee Member Arthur Goven, Chair of the Department of
Biological Sciences Costas Tsatsoulis, Interim Dean of the
Toulouse Graduate School
IMPROVING PROCESSING EFFICIENCY FOR FORENSIC DNA SAMPLES
Catherine Cupples Connon, B.S., M.S.
Dissertation Prepared for the Degree of
DOCTOR OF PHILOSOPHY
UNIVERSITY OF NORTH TEXAS
May 2015
Connon, Catherine Cupples. Improving Processing Efficiency for Forensic DNA Samples.
Doctor of Philosophy (Biology), May 2015, 238 pp., 54 tables, 65 figures, references, 90 titles.
The goal of this project was to reduce processing time for forensic DNA testing without
incurring significant added costs and/or the need for new instrumentation, while still
generating high quality profiles. This was accomplished by: 1) extraction normalization using
the ChargeSwitch® Forensic DNA Purification Kit such that a small range of DNA concentrations
was consistently obtained, eliminating the need for sample quantification and dilution; 2)
developing fast PCR protocols for STR primer sets using shorter amplification methods, low
volume reactions and non-fast thermal cyclers; and 3) developing a quicker 3130xl Genetic
Analyzer detection method using an alternative polymer/array length combination. Extraction
normalization was achieved through a reduction in bead quantity, thereby forcing an increase
in bead binding efficiency. Four products (AmpliTaq Gold® Fast PCR Master Mix, KAPA2G™ Fast
Multiplex PCR Kit, SpeedSTAR™ HS DNA Polymerase and Type-it Microsatellite PCR Kit) were
evaluated for low volume (3μl) fast PCR on a 384-well Veriti® thermal cycler with the Identifiler
primer set. KAPA2G™ was selected for 3μl fast PCR protocols using PowerPlex 16 HS and
Identifiler Plus primer sets (42-51min), as well as 5μl and 6μl Identifiler fast reactions on a 9700
thermal cycler (51-60min). Alternative detection (POP-6™/22cm) achieved 24-28min run times,
but with decreased resolution as compared to traditional POP-4®/36cm detection for alleles
>200bp; however, 1bp resolution was still obtainable for alleles <300bp. These modifications
resulted in robust databasing processes with up to a 37% reduction in processing time for
buccal swabs and Buccal DNA Collectors™ using the three primer sets evaluated (3μl fast PCR
reactions) and generated high quality STR profiles with ≥90% pass rates.
Copyright 2015
by
Catherine Cupples Connon
ii
ACKNOWLEDGEMENTS
I would like to thank Cellmark Forensics, A LabCorp Speciality Testing Group, for funding
a significant portion of this research project. I would like to further thank Cellmark’s Director of
Operations, Dr. Laura Gahn, not only for serving on my committee, but also for her, and Dr.
Aaron LeFebvre’s, willingness to work with my busy schedule. Thank you to Adam Hamburger
and Sheri Ayers for sharing in my excitement as this project progressed. Meghan Clement,
Tracey Cruz (of VCU) and Cynthia Smitherman deserve an additional word of thanks for their
continued guidance and much appreciated encouragement. Cynthia, you possess a wealth of
knowledge that is unsurpassed by any other that I know, and I can always count on you to shed
light on any topic – you are truly brilliant!
Several vendors provided fast polymerases, including Life Technologies, Kapa
Biosystems, ClonTech/Takara Bio and QIAGEN. Phillip Czar of Life Technologies went above and
beyond my expectations by generously donating numerous reagents. Thank you again, Phillip!
Dr. Benjamin, I owe a special thank you not only for accepting me into your program as a
transfer student, but also for taking me seriously when I presented you with my graduation
time frame. It was a tight schedule, but it all worked out as planned!
And to my beloved husband, I thank you endlessly for supporting me yet again through
graduate school. Life threw us some curve balls along the way (a job transfer across the country
and a new baby), but you were always by my side. You kept me focused and somehow I got
everything done just as planned!
iii
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS ..................................................................................................................... iii
LIST OF TABLES ................................................................................................................................ vi
LIST OF FIGURES .............................................................................................................................. ix
LIST OF ABBREVIATIONS ............................................................................................................... xiii
CHAPTER 1 INTRODUCTION ............................................................................................................ 1
CHAPTER 2 MATERIALS AND METHODS ....................................................................................... 39
2.1 Project Overview ................................................................................................... 39
2.2 Development of a Quicker Capillary Electrophoresis Detection Method for Forensic STR Profiles ............................................................................................. 40
2.3 Development of Fast PCR Protocols for Identifiler, Identifiler Plus and PowerPlex 16 HS ................................................................................................... 48
2.4 Development of a Normalized Extraction using the ChargeSwitch® Forensic DNA Purification Kit .............................................................................................. 68
CHAPTER 3 RESULTS AND DISCUSSION......................................................................................... 77
3.1 Development of a Quicker Capillary Electrophoresis Detection Method for Forensic STR Profiles ............................................................................................. 77
3.2 Development of Fast PCR Protocols for Identifiler, Identifiler Plus and PowerPlex 16 HS ................................................................................................. 106
3.3 Development of a Normalized Extraction using the ChargeSwitch® Forensic DNA Purification Kit ............................................................................................ 162
CHAPTER 4 CONCLUSIONS .......................................................................................................... 183
4.1 Improvements to Processing Efficiency .............................................................. 183
4.2 Future Studies ..................................................................................................... 193
iv
APPENDIX A SUPPLEMENTAL MATERIAL – RUN MODULES USED FOR DEVELOPMENT OF A QUICKER CAPILLARY ELECTROPHORESIS DETECTION METHOD ................................................. 197
APPENDIX B SUPPLEMENTAL MATERIAL – AMPLIFICATION PROTOCOLS USED FOR DEVELOPMENT OF VARIOUS FAST PCR METHODS ..................................................................... 203
APPENDIX C SUPPLEMENTAL MATERIAL – PARAMETERS FOR NORMALIZED EXTRACTION DEVELOPMENT ............................................................................................................................ 214
APPENDIX D SUPPLEMENTAL MATERIAL – RESULTS .................................................................. 217
REFERENCES ............................................................................................................................. 231
v
LIST OF TABLES
Page
1.1 Overview of Some DNA Extraction Chemistry and Platform Combinations ...................... 6
1.2 Controlling DNA Concentrations via Quantity of Beads ................................................... 11
1.3 Typical Thermal Cycling Protocols for Standard and Fast PCR ......................................... 18
1.4 Products Selected for Fast PCR Evaluation with the Identifiler Primer Set ...................... 21
1.5 Added Costs per Reaction for Fast Products .................................................................... 22
1.6 Advertised Maximum Ramp Rates (°C/sec) for Various Thermal Cyclers ........................ 23
1.7 Recommended Polymer Type and Array Length Combinations ....................................... 31
1.8 Parameters for the Default HID Run Module ................................................................... 33
2.1 Preparation of Amplification Product for Capillary Electrophoresis Detection ............... 41
2.2 Pass/Fail Detection Guidelines ......................................................................................... 45
2.3 Fast PCR Reaction Composition for Initial Evaluation ...................................................... 49
2.4 Fast PCR Thermal Cycling Parameters for Initial Evaluation............................................. 49
2.5 Final Fast PCR Reaction Compositions Compared to Standard PCR ................................. 55
2.6 Final Fast PCR Thermal Cycling Parameters Compared to Standard PCR......................... 55
2.7 First Pass Analysis Guidelines ........................................................................................... 62
3.1 Comparison of Traditional and Alternative Detection Methods ...................................... 83
3.2 Effects of Run Voltage, Run Temperature and Current Stability on ILS Sizing Quality .... 93
3.3 Effects of Run Voltage and Run Temperature on ILS Sizing Quality and Precision .......... 96
3.4 Comparison Between POP-4 and POP-6 Detection ........................................................ 103
3.5 Summary of Validated POP-6/22cm Injection Times ..................................................... 105
vi
3.6 Average Percent Stutter for n+4 and n-8 Stutter ........................................................... 152
3.7 Profile Summary From Large Data Set ............................................................................ 153
3.8 Lot-to-Lot Variation of KAPA2G™ Fast Multiplex PCR Kit ............................................... 155
3.9 Variation Due to Storage Conditions of KAPA2G™ Fast Multiplex PCR Kit .................... 156
3.10 Summary of DNA Recoveries .......................................................................................... 163
3.11 Sensitivity Range for Normalized Extraction and Fast PCR............................................. 171
3.12 Reproducibility of DNA Concentrations Within and Between Normalized Extraction Batches ............................................................................................................................ 172
3.13 Differences Between Duplicates Within and Between Normalized Extraction Batches ............................................................................................................................ 172
3.14 Bead Binding Capacities .................................................................................................. 173
3.15 Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: High Quality Swabs ....................................................................................... 174
3.16 Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: Old, Degraded Quality Swabs ....................................................................... 179
3.17 Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: Punches ......................................................................................................... 181
4.1 Time and Cost Savings for New First Pass Processes ...................................................... 185
A.1 Default Spectral Calibration Run Modules ..................................................................... 198
A.2 Spectral Calibration Run Modules Evaluated With Alternative Polymer/Array Combinations .................................................................................................................. 198
A.3 Default and Established Run Modules ............................................................................ 199
A.4 Run Modules Evaluated With NanoPOP4/36cm Array ................................................... 199
A.5 Initial Run Modules Evaluated With POP-6/22cm Array ................................................ 200
A.6 Run Modules Evaluated With POP-6/22cm Array for Peak Height and Pull-up Reduction ....................................................................................................................... 200
vii
A.7 Run Modules Evaluated With POP-6/22cm Array for ILS Sizing Quality Improvements ................................................................................................................. 201
A.8 Run Modules Evaluated With POP-6/22cm Array for Precision Improvements ............ 201
A.9 Validated POP-6/22cm Array Run Modules ................................................................... 202
B.1 Thermal Cycling Parameters for Standard PCR .............................................................. 204
B.2 Thermal Cycling Parameters for 3μl Identifiler Fast PCR Development With KAPA2G ........................................................................................................................... 204
B.3 Thermal Cycling Parameters for 3μl Identifiler Fast PCR Development With AmpliTaq Gold Fast ......................................................................................................................... 205
B.4 Thermal Cycling Parameters for 3μl Identifiler Fast PCR Development With SpeedSTAR ...................................................................................................................... 206
B.5 Thermal Cycling Parameters for 3μl Identifiler Fast PCR Development With Type-it .... 207
B.6 Thermal Cycling Parameters for 5μl Identifiler Fast PCR Development With KAPA2G .. 208
B.7 Thermal Cycling Parameters for 3μl Identifiler Plus Fast PCR Development With KAPA2G ........................................................................................................................... 208
B.8 Thermal Cycling Parameters for 3μl PowerPlex 16 HS Fast PCR Development With KAPA2G ........................................................................................................................... 209
B.9 Validated Thermal Cycling Parameters for 3μl, 5μl and 6μl Fast PCR Protocols With KAPA2G ........................................................................................................................... 210
B.10 CPALS Values to Describe Preferential Amplification Based on Locus Size .................... 213
C.1 Evaluated Parameters for Normalized Extraction Development ................................... 215
C.2 Validated Normalized ChargeSwitch Extraction ............................................................. 216
viii
LIST OF FIGURES
Page
1.1 Current Cellmark Forensics “first pass” process for databasing reference samples .......... 5
1.2 Reducing amplification time through ramp rates ............................................................ 24
1.3 One base-pair resolution .................................................................................................. 32
2.1 Positioning of allelic ladders ............................................................................................. 44
2.2 Positioning of allelic ladders for validation ....................................................................... 47
2.3 Location of swab cuttings ................................................................................................. 69
2.4 Buccal punch (6mm) and swab cutting locations (optimization) ..................................... 71
2.5 Buccal punch (6mm) and swab cutting locations (validation) .......................................... 72
3.1 Spectral calibrations obtained using NanoPOP4/36cm array with default POP-4/36cm array modules ................................................................................................................... 79
3.2 Spectral calibrations obtained using NanoPOP4/36cm array with modified run modules............................................................................................................................. 79
3.3 Identifiler Plus profile obtained using NanoPOP4/36cm array ........................................ 80
3.4 Internal size standards using traditional POP-4/36cm and alternative POP-6/22cm detection ........................................................................................................................... 84
3.5 Samples exhibiting first injection effect ........................................................................... 85
3.6 Resolution using traditional and alternative detection methods .................................... 87
3.7 Detectable pull-up peaks from traditional and alternative detection methods .............. 89
3.8 Frequency and percent pull-up from traditional and alternative detection methods ..... 91
3.9 Peak heights from traditional and alterative detection methods .................................... 91
3.10 Allele sizing differences for traditional and alternative detection methods .................... 94
3.11 Resolving power of traditional and alternative detection methods .............................. 100
3.12 ILS migration for traditional and alternative detection methods .................................. 101
ix
3.13 Allele sizing differences for Identifiler Plus and PowerPlex 16 HS ................................. 104
3.14 Representative profiles from initial evaluation of four fast PCR methods .................... 109
3.15 Effects of annealing temperature on inter-locus balance for Identifiler fast PCR with KAPA2G ........................................................................................................................... 111
3.16 Sensitivity of standard Identifiler and four fast PCR protocols ...................................... 128
3.17 Peak height summary for standard Identifiler and four fast PCR protocols .................. 128
3.18 Artifacts for standard Identifiler and four fast PCR protocols ........................................ 129
3.19 Optimal DNA input ranges for standard Identifiler and four fast PCR protocols ........... 130
3.20 Stochastic thresholds for standard Identifiler and four fast PCR protocols ................... 132
3.21 Precision of allele sizing for standard Identifiler and four fast PCR protocols ............... 133
3.22 Percent stutter (n-4) for standard Identifiler and four fast PCR protocols .................... 134
3.23 Sensitivity of fast PCR protocols using KAPA2G .............................................................. 142
3.24 Peak height summary for fast PCR protocols using KAPA2G .......................................... 143
3.25 Artifacts for fast PCR protocols using KAPA2G ............................................................... 144
3.26 Optimal DNA input ranges for fast PCR protocols using KAPA2G .................................. 146
3.27 Stochastic thresholds for fast PCR protocols using KAPA2G .......................................... 147
3.28 Precision of allele sizing for fast PCR protocols using KAPA2G ...................................... 148
3.29 Average percent stutter (n-4) for fast PCR protocols using KAPA2G ............................. 149
3.30 Maximum observed stutter (n-4) for fast PCR protocols using KAPA2G compared to standard PCR ................................................................................................................... 151
3.31 Fast PCR artifact at Amelogenin ..................................................................................... 154
3.32 Effect of storage conditions of peak height .................................................................... 157
3.33 Effect of annealing/extension temperature on non-specific amplification ................... 158
3.34 Effect of final extension length on -A and +A ................................................................. 160
x
3.35 Effects of final extension length and input DNA on profile quality ................................ 160
3.36 Effects of annealing/extension temperature and final extension on profile quality ..... 161
3.37 Tested sensitivity ranges for normalized extraction coupled with fast PCR .................. 169
3.38 Percent full profiles obtained using normalized extraction coupled with fast PCR ....... 170
3.39 Detection of non-failing artifacts using current and normalized extractions ................ 177
4.1 Comparison of new and current Cellmark Forensics “first pass” processes for databasing reference samples ........................................................................................ 184
B.1 Three models for preferential amplification based on locus size .................................. 211
B.2 CPALS values for three models of preferential amplification ......................................... 212
D.1 Precision of allele sizing for traditional and alternative detection methods ................. 218
D.2 One base-pair resolution using traditional and alternative detection methods ............ 219
D.3 Representative Identifiler Plus allelic ladders from traditional and alternative detection methods .......................................................................................................... 220
D.4 Representative PowerPlex 16 allelic ladders from traditional and alternative detection methods .......................................................................................................... 221
D.5 Representative Identifiler Plus profiles from traditional and alternative detection methods .......................................................................................................................... 222
D.6 Representative PowerPlex 16 profiles from traditional and alternative detection methods .......................................................................................................................... 222
D.7 Representative 3μl Identifler profiles from fast and standard PCR ............................... 223
D.8 Representative Identifiler profiles from fast and standard PCR..................................... 224
D.9 Representative Identifiler Plus profiles from fast and standard PCR ............................. 225
D.10 Representative PowerPlex 16 HS profiles from fast and standard PCR ......................... 225
D.11 Representative profiles obtained from new and current 3μl PowerPlex 16 HS first pass options .................................................................................................................... 226
D.12 Representative profiles obtained from new and current 3μl Identifiler Plus first pass options ............................................................................................................................ 227
xi
D.13 Representative profiles obtained from new and current 3μl Identifiler first pass options ............................................................................................................................ 228
D.14 Representative profiles obtained from new and current 6μl Identifiler Plus first pass options ............................................................................................................................ 229
D.15 One base-pair resolution for large sized alleles using traditional detection .................. 230
xii
LIST OF ABBREVIATIONS
-A Minus A, incomplete adenylation
AG AmpliTaq Gold® Fast PCR Master Mix
bp Base-pair
CI Confidence interval
CODIS Combined DNA Index System
CV Coefficient of variation
DNA Deoxyribonucleic acid
HS PowerPlex® 16 HS System
HSF PowerPlex® 16 HS System, fast PCR
HSS PowerPlex® 16 HS System, standard PCR
ID AmpFℓSTR® Identifiler® PCR Amplification Kit
ID+ AmpFℓSTR® Identifiler® Plus PCR Amplification Kit
ID+F Identifiler® Plus primer set, fast PCR
ID+S AmpFℓSTR® Identifiler® Plus PCR Amplification Kit, standard PCR
IDF Identifiler® primer, fast PCR
IDS AmpFℓSTR® Identifiler® PCR Amplification Kit, standard PCR
ILS Internal lane/size standard
IT Information technology
KP KAPA2G™ Fast Multiplex PCR Kit, KAPA2G™ Fast Multiplex Mix
LOR Loss of resolution
LPH Locus peak height, sum of
xiii
NSA Non-specific amplification
OL Off ladder allele
PCR Polymerase chain reaction
PH Peak height
RFLP Restriction fragment length polymorphism
rfu Relative fluorescence units
SS SpeedSTAR™ HS DNA Polymerase
STR Short tandem repeat
TI Type-it® Microsatellite PCR Kit
TPH Total sum of peak heights (for profile)
xiv
1
CHAPTER 1
INTRODUCTION
Time and cost are major concerns for nearly everything, especially with respect to the
scientific community; forensics is not a stranger to this concept either. In the forensics world,
results are always needed faster, and at a lower cost, to boot. In the past (1980-1990s), it took
several weeks, if not months, to process samples for genetic profiling using RFLP technology
(Butler, 2012, p. 104). Today, a forensic sample can easily be processed in a matter of days with
standard STR technology, and even within 24 hours if necessary (Butler, 2012, p. 104). Still, one
can ask, “Can results be obtained even quicker than that?” And if so, would it be possible to
significantly improve processing time for all workflow – not just the rush or high priority cases –
and to simultaneously reduce processing time/costs for more routine casework as well?
Currently, there are a number of reasons why it would be beneficial to obtain DNA profiling
results in a more timely fashion, including the necessity to increase sample throughput either
due to an increase in incoming samples or to reduce a backlog (Hedman et al., 2008), as well as
the need to report results as soon as possible for a high profile case or for investigative leads
associated with a detained person of interest (Gangano, Elliot, Anoruo, Gass, & Buscaino, 2013).
Direct amplification and rapid DNA analysis instruments have been utilized for these purposes
(Butler, 2012, p. 40; Gangano, Elliot, Anoruo, Gass, & Buscaino, 2013; Gray, Crowle, & Scott,
2014; LaRue, Moore, King, Marshall, & Budowle, 2014; Park, Kim, Yang, & Lee, 2008). However,
both methods have associated drawbacks. Direct amplification can face PCR inhibitition due to
components found in a sample (e.g., heme in whole blood, dirt or fabric dyes), as well as over-
or under-amplification due to the narrow range of input DNA amounts allowed by STR multiplex
2
kits (Verheij, Harteveld, & Sijen, 2012). On the other hand, rapid instruments are extremely
expensive (acquistion, maintenance and supplies) and comparitively low throughput, and thus
are not cost effective for high throughput applications (D. Murphy, personal communication,
April 15-17, 2014). Although, the majority of this chapter focuses on ways to reduce processing
time, cost reduction is discussed at a later point in this dissertation.
In order to improve sample processing time, there are three basic approaches that can
be taken: 1) a completely new method can be developed that yields a new type of genetic
profile, 2) a new method can be developed that yields a genetic profile similar to the one
currently in use (or at least can be compared to current ones), 3) or an existing method can be
improved upon in a fashion that would still yield the same type of genetic profile. Forensic DNA
technology is anything but static and, at any given time, current techniques are regularly being
improved upon (Foster & Laurin, 2012; Giese, Lam, Selden, & Tan, 2009; Greenspoon et al.,
2004; Leclair et al., 2003; Westring, Kristinsson, Gilbert, & Danielson, 2007). Additionally, many
new and novel techniques are continuously being explored/developed (Daniel et al., 2015;
Fordyce et al., 2015; Gangano et al., 2013; Lennard Richard et al., 2012; Wang, Luo, Pan, Liao, &
Hou, 2012; Wasserstrom et al., 2013). Operating within this ever changing niche in science can
be difficult for crime laboratories and the criminal justice system because process changes
require training, document and IT updates, and depending upon the situation, possibly changes
to the databases that house genetic profiles. Smaller changes are generally preferred over
larger ones because they are quicker and easier to implement. An example of a major change
would be the implementation of the presently used nuclear autosomal STR profiles, which
immediately rendered useless all profiles in existing RFLP databases. Today, CODIS databases
3
that house DNA profiles consisting of allelic information for only 13 loci (i.e., the CODIS core
loci) cannot be compared to genetic profiles derived from the application of other technologies
(e.g., Y-STRs, SNPs, microRNAs, mRNAs, other non-CODIS loci). Thus, though meaningful for
their individual applications, a forensic scientist would not likely perform one of these tests as a
substitute for another that would yield a profile of the CODIS core loci. Given the existing
foundation based upon these 13 loci, it would be difficult to convert to a new type of genetic
profile that fails to collect this information all together; however, it would be easier for CODIS
to add additional loci to supplement the core 13. This exact idea has already been proposed,
and the FBI is presently working towards expanding the CODIS core loci to include a total of 20
STR loci (Ge, Eisenberg, & Budowle, 2012).
Keeping implementation difficulties in mind, it seems more viable to take an existing
process or method and modify it to improve processing time. Therefore, the goal of this project
was to select a robust process utilized at an existing DNA crime laboratory, which generates
DNA profiles amplified from primer sets that are widely accepted in the forensic community,
and improve upon the processing efficiency via modifications to that process.
Sample workflow in a databasing facility (i.e., one that processes a large number of high
quality reference samples) tends to be faster paced, higher throughput and more uniform than
that of a casework facility (i.e., one that processes a variety of evidentiary samples using various
protocols), making it (databasing) a good candidate for process improvement. In a high
throughput environment, many profile quality deficiencies that might arise related to
uncommon or rare genotypes (e.g., tri-alleles, weak tri-alleles, microvariants, resolution of
alleles differing by one base-pair, etc.) should be more quickly identified given the sheer
4
number of samples processed in a relatively short period of time. Additionally, reference
samples (including buccal swabs, buccal collectors and bloodstain cards) are high quality, single
source and readily available, making them excellent sample sources for process improvement
as compared to casework samples, which tend to be much more diverse and of poorer quality.
Thus, it is advantageous to improve the testing process for reference samples in a databasing
environment first, and then transition those or similar improvements onto casework samples
after the new process has proven to be highly successful in the databasing setting.
The general process for DNA profiling consists of numerous different steps:
accessioning, sample cutting, DNA extraction, DNA quantification, pre-amplification dilution/
sample normalization, STR amplification, STR detection and STR profile analysis. Figure 1.1
illustrates the typical amount of time spent on each of these tasks using current procedures at
the DNA crime laboratory selected for process improvement using a databasing batch
consisting of 89 reference samples (plus one reagent blank, two positive amplification controls,
one negative amplification control and three allelic ladders), totaling approximately 14.25
hours. From this breakdown, it is clear that extraction, amplification and profile detection
account for substantial portions of the overall process (14-35%), and any proportional
reduction in time spent on these tasks would greatly reduce total processing time.
In addition to the gold standard manual organic extraction procedure (Sambrook,
Fritsch, & Maniatis, 1989), there are a wide variety of DNA extraction chemistry and platform
combinations commercially available to forensic laboratories (see Table 1.1) (Aebischer, Beer, &
Hoffman, 2014; Anslinger, Bayer, Rolf, Keil, & Eisenmenger, 2005; Davis, King, Budowle,
5
Eisenberg, & Turnbough, 2012; Frégeau, Lett, & Fourney, 2010; Greenspoon et al., 2004;
Phillips, McCallum, & Welch, 2012; Witt, Neumann, Zierdt, Gébel, & Röscheisen, 2012).
Figure 1.1. Current Cellmark Forensics “first pass” process for databasing reference samples. For each step, time spent is displayed in hours and percent of the entire process. Samples are manually accessioned and cut into a 96-well plate, followed by DNA extraction using a BioSprint 96 (QIAGEN, Valencia, CA) or KingFisher® 96 (KingFisher 96; Thermo Scientific, Vantaa, Finland) instrument and the ChargeSwitch® Forensic DNA Purification Kit (ChargeSwitch; Life Technologies, Grand Island, NY). As recommended by the manufacturer of the ChargeSwitch kit, samples are then quantified using the Quant-iT™ PicoGreen® dsDNA Quantitation Kit coupled with the Quant-iT™ PicoGreen® dsDNA Quantitation Reagent (PicoGreen; Life Technologies) (Invitrogen™ Life Technologies, 2005) and a FLUOstar microplate reader (BMG LABTECH, Ortenberg, Germany), followed by a pre-amplification dilution to normalize samples for amplification. Next, samples are amplified using one of several commercially available STR amplification kits in a low volume and/or reduced cycle reaction (3µl or 6µl). STR profiles are then detected using a 3130xl Genetic Analyzer (3130xl; Life Technologies) equipped with POP-4® Polymer (POP-4; Life Technologies) and 36cm array. Lastly, STR profiles are analyzed using GeneMapper® ID v3.2.
1.0hr 7.0%
0.50hr 3.5%
2.0hr 14%
0.75hr 5.3%
0.50hr 3.5%
3.3hr 23%
5.0hr 35%
1.3hr 8.8%
Accessioning
Cut
Extraction
Quantification
Dilution
Amplification
Detection
Analysis
6
Table 1.1
Overview of Some DNA Extraction Chemistry and Platform Combinations
Extraction Chemistry Manual Low Throughput
Automated Instrument
High Throughput Automated Instrument
BioSprint DNA Blood Kita No BioSprint 15a BioSprint 96a,b
DNA Investigator Kita Yes EZ1® instrumentsa,
QIAcubea TECAN
Instruments
ChargeSwitch® Forensic DNA Purification Kitc
Yes iPrep™ Purification
Instrumentc
BioSprint 96a,b, TECAN
Instruments PrepFiler® Forensic DNA
Extraction Kitc Yes Automate Express™ c
TECAN instruments
DNA IQ™ Systemd Yes Maxwell® 16d BioMek® 2000,
TECAN workstations
Note. Some of the various extraction chemistry and platform combinations that available for forensic use are listed above. aManufactured by QIAGEN.
bThe BioSprint 96 and KingFisher 96 instruments are nearly identical and can be
interchanged. cManufactured by Life Technologies.
dManufactured by Promega.
eManufactured by Beckman
Coulter, Brea, CA.
Automated DNA extraction platforms typically utilize a 96-well format for high throughput
processing on instruments such as TECAN robotic liquid handling workstations, BioSprint 96 and
KingFisher 96 or on lower throughput instruments (≤16 samples per extraction run) such as
BioSprint 15 (QIAGEN), QIAcube (QIAGEN), EZ1® instruments (EZ1; QIAGEN), iPrep™ Purification
Instrument (iPrep; Life Technologies), Automate Express™ (Life Technologies) and Maxwell® 16
Forensic Instrument (Maxwell 16; Promega, Madison, WI).
The scope of this study was not to evaluate different extraction procedures or
platforms, but to increase processing efficiency via a reduction in processing time for an
existing process within an actual forensic DNA laboratory. Thus, though various extraction
options were available, this study utilized the extraction procedure that was employed by
7
Cellmark – a ChargeSwitch extraction via purification on a BioSprint 96/KingFisher 96
instrument. The ChargeSwitch chemistry is a “magnetic bead-based technology that utilizes a
switchable surface charge dependent upon the pH of the surrounding buffer to facilitate nucleic
acid purification” (Invitrogen™ Life Technologies, 2005). In low pH environments, the beads
have a positive charge and bind the negatively charged DNA backbone. Two wash steps are
utilized to remove proteins and other contaminants, neither of which are bound by the beads.
Raising the pH to 8.5 using a low salt elution buffer neutralizes the previously charged bead
surface, thereby releasing the DNA into the elution buffer.
Although the ChargeSwitch chemistry was not originally developed to be used the
BioSprint 96 or KingFisher 96, Cellmark utilizes these for DNA extractions. The BioSprint 96 and
KingFisher 96 are manufactured by different vendors, but are nearly identical, are used
interchangeably at Cellmark with the ChargeSwitch chemistry and are referred to as a magnetic
particle separator (MPS) in this dissertation. Following a brief incubation (minimum of one
hour), samples can be purified on an MPS in about 25 minutes, which is comparable to the
length of extraction on a lower throughput instrument (the iPrep) developed specifically for use
with the ChargeSwitch chemistry. I reviewed Cellmark’s ChargeSwitch extraction procedure and
determined that it was already streamlined under their conditions (short of reducing incubation
time to less than one hour). Therefore, I decided to investigate the possibility of modifying the
extraction procedure, such that it would yield a consistent sample DNA concentration via the
extraction process itself. Samples would thus proceed from extraction to amplification, without
having to quantify and adjust the concentration to the required amount. This is an acceptable
process flow for reference samples, as they are not required to undergo a human-specific
8
quantification step prior to amplification (FBI, 2011). Additionally, this proposed process would
be similar to some direct amplification methods (sample substrate is briefly incubated in a
buffer, followed by direct amplification of a small aliquot of the buffer (Park et al., 2008)), but
would aim to produce higher quality extracts than those used in direct amplification and,
ultimately, higher quality STR profiles.
Promega and Life Technologies manufacture multiplex direct amplification kits, such as
the PowerPlex® 18D System (PowerPlex 18D; Promega) and the AmpFℓSTR® Identifiler® Direct
PCR Amplification Kit (Identifiler Direct; Life Technologies), respectively, and both are optimized
for amplification of untreated FTA® (Whatman) punches (Life Technologies, 2012; Promega,
2014b). Myers, King, and Budowle (2012) evaluated both of these direct kits using buccal cells
on FTA® paper and obtained “first pass” success rates of 96.25% (6% of profiles exhibiting
saturation) for PowerPlex 18D and 95% (3.3% of profiles exhibiting saturation) for Identifiler
Direct. Myers et al. also noted that when attempting to setup direct amplification with
AmpFℓSTR® Identifiler® Plus PCR Amplification Kit (Identifiler® Plus; Life Technologies), 15% of
the profiles exhibited minus A (-A), or incomplete adenylation, which could not be overcome.
Other groups have developed direct amplification methods using common forensic
primer sets for casework type samples and reference type samples. Park et al. (2008) used a
direct PCR buffer (AnyDirect™ PCR buffer; BioQuest Inc., San Francisco, CA) in conjunction with
AmpFℓSTR® Identifiler® PCR Amplification Kit (Identifiler; Life Technologies), AmpFℓSTR®
Yfiler™ PCR Amplification Kit (Yfiler; Life Technologies) and the PowerPlex® Y System
(PowerPlex Y; Promega) for blood and saliva (spotted on FTA® paper) reference samples. Full
profiles were obtained from 100% (n=50) of the blood spots, but some peak height ratios (PHR)
9
were less than 0.40 for both sample types, and some saliva samples exhibited poor
amplification, including allelic dropout (Park et al., 2008). Gray et al. (2014) demonstrated that
direct amplification (<3 hours) was possible for casework type samples (i.e., blood stains and
blood swabs) using the PowerPlex® 21 System (PowerPlex 21; Promega) with 80% of samples
yielding full profiles. Verheij et al. (2012) demonstrated a unique start to finish process in under
six hours (for 19 forensically relevant samples) involving direct and fast amplification with
Phusion® Flash High-Fidelity PCR Master Mix (Phusion Flash; Thermo Scientific), demonstrating
46-100% full profiles (depending on sample type) using the primer set from AmpFℓSTR® SGM
Plus™ (SGM+; Life Technologies).
Thus, many research groups have demonstrated that full profiles can be obtained from
FTA® punches using direct amplification with high success rates (Myers et al., 2012; Park et al.,
2008), but I am unaware of any studies involving normalized extraction to eliminate the
quantification and/or pre-amplification dilution steps. Normalized extraction and direct
amplification both reduce processing time by eliminating quantification/dilution, but
normalized extraction may prove to be more robust for a variety of sample types (or at least
provide a means to optimize slightly different normalized extraction procedures for different
sample types) given that direct amplification has been shown to produce non-optimal results
depending on the specific protocol. Furthermore, the existing overall process that I selected for
this study utilizes significantly reduced amplification volumes (3-6µl, compared to 25μl for
standard amplification) as a key cost-savings maneuver, and direct amplification at such low
volumes would likely be unsuccessful. Therefore, in order to maintain these low reaction
10
volumes, this particular process flow would more likely benefit from normalized extraction
rather than using a direct amplification approach.
The key to normalized extraction is limiting the maximum quantity of DNA that can
ultimately be extracted and purified, coupled with utilizing samples that contain at least that
amount of DNA, if not slightly more, such that similar DNA recoveries (i.e., the maximum
amount) can be obtained from all samples. Theoretically, ChargeSwitch, like other magnetic
bead-based extraction procedures (e.g., DNA IQ™ and PrepFiler®), should have an upper limit to
the amount of DNA that can be purified, which is directly related to the quantity of beads that
is added to each sample. DNA IQ™ was among the first of the commercially available bead-
based extraction procedures and heavily advertised this exact feature, indicating maximum
yields of 50-100ng for FTA® blood-card punches, 50-200ng for liquid blood and 100-500ng for
buccal swabs (Promega, 2013). Likewise, ChargeSwitch notes a maximum binding capacity of 5-
10µg genomic DNA per milligram of beads (Invitrogen™ Life Technologies, 2005), and maximum
DNA recovery should decrease as the bead quantity utilized decreases. Based upon this
information, it should be possible to control DNA recoveries via bead quantity and thus,
attempt normalization during the extraction procedure itself. However, it is important to note
that it is assumed that maximum DNA recoveries can only be achieved when there is an excess
of DNA present in the sample (Promega, 2013), and it is not guaranteed that this maximum will
be obtained for all samples, especially ones with low quantities of DNA. Previous studies have
shown that individuals shed varying amounts of skin and buccal cells (Frantz Burger, Song, &
Schumm, 2005; Lowe, Murray, Whitaker, Tully, & Gill, 2002). Therefore, when using bead
quantity to control yields, DNA recoveries could range anywhere from zero (from extremely
11
poor shedders) up to the maximum recoverable amount (dependent upon bead quantity and
bead binding capacity). Moreover, STR amplification requires a narrow range of DNA input
(typically 0.5-2.0ng) in order to produce quality profiles (Butler, 2012, p. 49). Extraction
normalization would have to take into account shedder status, as well as this targeted input
range, and should aim for a maximum DNA recovery that coincides with the upper end of the
optimal DNA input range.
The procedure under evaluation utilizes a quarter reaction of the standard ChargeSwitch
extraction (Invitrogen™ Life Technologies, 2005), in which 5µl of beads are added to each
sample, which is later recovered in a final elution volume of 60µl or 80µl, depending on sample
type (Buccal DNA Collector™ [Bode Technology, Lorton, VA] punches versus buccal swabs,
respectively). Given bead binding capacity (5-10µg DNA), bead concentration (25mg/ml) and
final elution volume (60µl), the predicted maximum DNA concentrations can be calculated
based upon the quantities of beads used (see Table 1.2).
Table 1.2
Controlling DNA Concentrations via Quantity of Beads
Volume (µl) of Beads Predicted Maximum DNA Concentration (ng/µl)
5.00 10.4-20.8 4.00 8.33-16.7 3.00 6.25-12.5 2.00 4.17-8.33 1.00 2.08-4.17
0.500 1.04-2.08 0.250 0.521-1.04
Note. Predicted maximums are based on samples with a final volume of 60.0µl. Minimum DNA concentrations can be as low as 0ng/µl.
12
In practice, the current extraction procedure under review has in fact yielded maximum
DNA concentrations of approximately 20ng/µl, which coincides with the maximum predicted
yields when using 5µl of beads. Thus, the predicted values in Table 1.2 are assumed to be
reasonably accurate, and it appears feasible to obtain DNA extracts with an appropriate DNA
concentration that can be used for STR amplification targeting ~0.5-2.0ng of input DNA by
reducing bead volume to <5µl. Furthermore, it is important to note that in the procedure under
evaluation, first pass amplification uses 3-6µl reaction volumes, allotting 1.2-2.4µl of DNA
isolate per reaction. Currently in this process, DNA is transferred from a 96-well dilution plate
to the amplification plate using the Nanodrop™ Express (BioNex Solutions Inc., Sunnyvale, CA),
which is calibrated to transfer volumes as small as 0.5µl (±10%). Therefore, the volume of DNA
added to the amplification reaction could potentially be reduced, if necessary, to as low as 0.5µl
To accommodate processing of higher concentrations of DNA. If this approach is pursued, then
water can be added to the master mix to bring reactions up to final volume.
In addition to controlling DNA recoveries via bead quantity, sample size (i.e., substrate
size) and elution volume may also be altered in order to obtain a tight DNA concentration
range. Increasing the size of the sample cutting should not impact the maximum recovery, but
could help to increase recoveries from samples with low quantities of DNA (e.g., from low
shedders). However, it remains to be seen whether DNA recovery will remain the same for
samples having DNA quantities in excess of the maximum recoverable amount even as sample
size increases. It is also unclear whether extract quality for samples with DNA in extreme excess
of the maximum recoverable amount will be equal to that obtained from samples with DNA
amounts close to or less than the maximum recoverable amount. It is possible that samples
13
with an extreme excess quantity of DNA may exhibit too much competition for the surface area
of the beads and, in doing so, may trap PCR inhibitors or other impurities and disrupt the
purification process. Conversely, DNA from samples with very low levels of total DNA may have
a difficult time coming into contact with and binding beads if bead quantity is substantially
reduced, as would be proposed by the calculations in Table 1.2, and/or may experience
competition with other substances for bead binding sites (Frégeau & De Moors, 2012).
As mentioned above, adjusting elution volume could be another way to change the final
DNA concentration of the extracted samples. Concentration and volume are inversely
proportional when quantity (e.g., amount of DNA) is held constant. Thus, by applying this basic
principle, DNA concentration will increase if the elution volume decreases and vice versa.
Elution volume can thus be modified to fine tune the resulting range of DNA concentrations.
However, in the current process, samples are eluted in an elution buffer (pH=8.5) supplied with
the ChargeSwitch kit and diluted with water prior to amplification. It is unknown whether PCR
inhibitors are present in the undiluted extract itself, but are subsequently diluted to an
acceptable level during the dilution process of the current first pass procedure. Furthermore,
once modifications are made to the extraction procedure to normalize resulting DNA
concentrations, it is unknown whether these changes will alter the amount of PCR inhibitors (if
any at all) that is carried over into the sample extract. If there is an increase in inhibitors,
amplification of undiluted extract may result in PCR inhibition. Additionally, even if the amount
of PCR inhibitors does not increase as a result of proposed modifications to the extraction
procedure, the level of inhibitors may be sufficient to cause PCR inhibition simply because they
are no longer being “diluted out.” If either of these occurs, PCR inhibition issues may be
14
resolved by reducing the final elution volume, such that the extracts have a higher DNA
concentration, thus requiring less volume DNA for amplification and allowing the PCR inhibitors
to be diluted via the addition of water to the master mix at the time of amplification.
Ultimately, the critical issue is that of inhibitor:DNA ratio and finding the right balance to obtain
the same quantity of DNA that would be used in the current process without increasing the
level of inhibitors to the point of PCR inhibition in the proposed process.
Furthermore, a particular extraction procedure can behave differently for different
sample types. It would be beneficial for the modified process to be robust for a range of sample
types. Again, the scope of this project is to focus on databasing reference samples, with the
intention of expanding it to forensic casework type samples in the future. Buccal swabs and
blood are among the most commonly encountered databasing reference sample types (Butler,
2012, p. 8), but even these can be subdivided further in regard to the tool used for sample
collection. DNA from buccal cells can be collected and/or stored on various product styles from
numerous vendors, including sterile cotton tipped applicators (i.e., cotton swabs), sterile foam
tipped applicators, Buccal DNA Collector™, FTA® cards (Whatman®), buccal brushes, etc.
(Butler, 2012, p. 8; Frantz Burger et al., 2005; Hansen, Simonsen, Nielsen, & Hundrup, 2007;
Parsons & Bright, 2012; Richards et al., 1993; Walsh & Corey et al., 1992). Blood can be stored
on a stain card for future processing or can be processed as whole blood (Halsall et al., 2008;
Hansen et al., 2007). It has become evident that though these all are acceptable products and
modes of collection for reference samples, some may result in lower quantity and/or quality
extracts because of the sample and/or substrate itself and/or the extraction procedure used.
For epidemiologic studies, whole saliva samples are a good alternative to blood (Verheij et al.,
15
2012), both of which demonstrated superior results over buccal cells collected on swabs or
foam-tipped applicators (Hansen et al., 2007). On the other hand, higher success rates have
been obtained from blood samples compared to saliva transferred to FTA® indicating cards
(Park et al., 2008). Genomic DNA quality, as evaluated via 260/280nm and 260/230nm ratios
from UV light absorption, indicated that mouth swabs (i.e., buccal swabs) and oral samples
transferred to FTA® indicating cards are contaminated with carbohydrates, peptides, phenols,
buffer salts or other aromatic compounds, all of which may negatively impact PCR amplification
(Hansen et al., 2007). Additionally, the DNA IQ™ System has been shown to recover lower
quantities of DNA from Curity® (Covidien, Mansfield, MA), Kendall (now part of Covidien) and
Puritan® (Guilford, ME) brand cotton swabs (Puritan®, 2013; Wisconsin State Crime Laboratory,
2013). More well-known is the fact that blood contains heme and, if heme is not effectively
removed during the extraction/purification process, it can cause PCR inhibition (Verheij et al.,
2012). Using too large of a blood sample may reduce a procedure’s ability to remove heme as
well (Invitrogen™ Life Technologies, 2005). As a result, some commercially available extraction
methods offer slightly different protocols based on sample type. For example, the EZ1® DNA
Investigator Kit outlines numerous pretreatments for different samples, including, but not
limited to, whole blood, dried blood, saliva and various forensic-type samples (QIAGEN, 2009a).
Similarly, the ChargeSwitch chemistry is available for blood, serum, buccal cells and forensic
type samples (Invitrogen™ Life Technologies, 2005).
Aside from the actual sample substrate, some collection tools offer a more controlled
method for sample size and/or amount of genetic material. For example, when cutting cotton
swabs, it is impossible to collect the same size cutting from the same location of the swab every
16
time. This can be influenced by the individual cutting the swab, the tool used for cutting
(scissors versus various scalpels), the positioning of the swab stick and the size of the swab
itself, of which the latter two can vary, presumably due to manufacturing processes (personal
observation). Coupling this with the fact that the amount of genetic material on a swab is donor
dependent, opens up the possibility that a larger than desired range of DNA concentrations
may be obtained from swabs no matter how the extraction procedure is modified in an attempt
to normalize recoveries. On the other hand, buccal collectors, such as the Buccal DNA
Collector™, FTA® indicating cards and blood stain cards all offer a slightly more controlled
method to process buccal or blood samples. Instead of cutting this type of sample, a hole-
punch can be used to obtain nearly the same size “cutting” every time, which in and of itself
will help to narrow the overall range of DNA recoveries. To further reduce this range, the punch
should be collected from the same area of the collector/card every time. Buccal collectors and
stain cards will still be subject to various amounts of genetic material due to the donor’s
shedder status or white blood cell count but, overall a tighter range of DNA concentrations may
be obtainable by eliminating sample “cutting” issues.
In addition to reducing processing time by normalizing DNA recoveries during the
extraction step such that quantification and pre-amplification dilution can be eliminated, the
time spent amplifying samples can be reduced by employing fast PCR methods. Fast PCR is a
somewhat new term and can refer to any amplification process that takes less time to perform
than standard PCR, but there is another definition that includes the use of a higher than
standard number of very short, inefficient cycles in order to obtain more amplified product in
17
the same amount of time (or less) than standard PCR (Kopf-Sill, 2001). For this project, I am
referring to the former, more general definition of fast PCR.
The typical non-fast PCR protocol utilized in forensics employs the hot-start AmpliTaq
Gold® DNA Polymerase (AmpliTaq Gold; Life Technologies) and a thermal cycling protocol
lasting about three hours (Applied Biosystems, 2006a; Butler, pp. 71, 90), such as that displayed
in Table 1.3. Fast PCR is achieved by shortening nearly all of these amplification times. For
example, AmpliTaq Gold requires “hot-start” activation at 95°C for about ten minutes, whereas
fast polymerases have shorter activation times (1-2 minutes) (Butler, 2012, p. 80; New England
BioLabs® Inc., 2014). Additionally, AmpliTaq Gold has an extension rate of approximately
1kb/min (New England BioLabs® Inc., 2014), compared to 2-4kb/min for fast polymerases (New
England BioLabs® Inc., 2014). The most commonly used thermal cycler in the forensic
community is the GeneAmp® PCR System 9700 (9700; Life Technologies), which is typically used
in “9600 Emulation Mode” with a ramp rate of 1°C/sec, whereas fast PCR often employs fast
instruments (or even the 9700) using faster ramp rates (Butler, 2012, pp. 75, 91; Vallone, Hill, &
Butler, 2008). In addition to utilizing a fast polymerase (and/or buffer) and/or fast PCR thermal
cycler, other means of achieving fast PCR include employing a lower reaction volume, using a
2-step rather than 3-step PCR cycle, using thin-walled tubes and/or increasing DNA input
(Foster & Laurin, 2012; Giese et al., 2009; Kopf-Sill, 2001; Laurin & Frégeau, 2012; Vallone et al.,
2008; Verheij et al., 2012; Wang, Chang, & Hennessy, 2009). When one or more of these
techniques are utilized, overall amplification time can be reduced by employing a reduced total
number of PCR cycles, shortening cycle lengths and/or shortening activation/final hold steps
(see Table 1.3). Each of these options is discussed in more detail below.
18
Table 1.3
Typical Thermal Cycling Protocol for Standard and Fast PCR
Description Number of Cycles Temperature (°C) Length
Standard
Fast
Hot-Start Activation 1 95 10min 1-2min Denaturation
Annealing Extension
28-30 95 60
70-72
1min 1min 1min
5-10sec 15sec 0-5sec
Final Extension 1 60a or 72b 30-60min 1-10min Hold 1 4 or 25 ∞ ∞
Note. A typical non-fast amplification protocol takes about three hours (Applied Biosystems, 2006a; Butler, Advanced Topics in Forensic DNA Typing: Methodology, 2012, pp. 71, 90), whereas typical fast PCR protocols utilizing a fast polymerase and a fast thermal cycler with or without a combined annealing-extension step can be achieved in less than 20 minutes (Butler, Advanced Topics in Forensic DNA Typing: Methodology, 2012, pp. 80, 91; Foster & Laurin, 2012; Giese, Lam, Selden, & Tan, 2009). aTypical final extension temperature used with standard 3-step PCR cycling.
bTypical final extension temperature
used with 2-step PCR cycling that employs a combined annealing-extension step.
Numerous groups have developed fast PCR protocols ranging from 19min to less than
two hours in length, but many efforts have resulted in lower quality profiles than those seen
using standard amplification, especially with regard to an increase in n-4 stutter, -A, non-
specific amplification products, allelic dropout, elevated baseline and inter-locus imbalance, as
well as non-human cross reactivity (Foster & Laurin, 2012; Giese et al., 2009; Laurin & Frégeau,
2012; Vallone et al., 2008; Verheij et al., 2012; Wang et al., 2009). These fast PCR protocols
utilized various primer sets (AmpFℓSTR® COfiler® [COfiler; Life Technologies, Grand Island, NY],
Identifiler, AmpFℓSTR® NGM™ [NGM; Life Technologies], AmpFℓSTR® Profiler Plus® [Profiler
Plus; Life Technologies], SGM+ and AmpFℓSTR® SEfiler Plus™ [SEfiler+; Life Technologies]) and
polymerases (AmpliTaq Gold® DNA Polymerase, AB77, AB95, AB-1, AB-3 [all from Life
Technologies], SpeedSTAR™ HS DNA Polymerase [SpeedSTAR; Takara Bio Inc., Shiga, Japan],
KOD Hot Start DNA Polymerase [KOD; EMD Millipore, Gibbstown, NJ], Phusion Flash, Phusion®
19
High-Fidelity DNA Polymerase [Phusion; New England BioLabs® Inc., Ipswich, MA] and
PyroStart™ Fast PCR Master Mix [PyroStart; Fermentas Inc., Glen Burnie, MD]), as well as fast
(Piko® Thermal Cycler, 24-well [Piko; Thermo Scientific], C1000™ [C1000; Bio-Rad Laboratories,
Mississauga, ON], Mastercycler® ep gradient S [Mastercycler; Eppendorf, Hamburg, Germany]
and a custom designed microfluidic biochip) and non-fast thermal cyclers (9700 and Veriti®
[Veriti; Life Technologies]), demonstrating the versatility of the concept.
As alluded to above, numerous fast PCR polymerases are commercially available, but
not all are necessarily suitable for STR amplification. Specifically, a polymerase/buffer
combination suitable for forensic STR amplification needs to be capable of multiplexing 16 or
more loci with amplicons ranging in size from ~100bp to 500bp (Butler, 2012; Life Technologies,
2014b; Promega, 2012; Promega, 2014c), all the while maintaining high fidelity. I preferred to
work with a polymerase that generated adenylated products due to the fact that commercially
available STR amplification kits yield adenylated products and these are compared to allelic
ladders comprised of adenylated alleles. Thus, maintaining the norm would be easier than
evaluating a polymerase that generates blunt end products because a new allelic ladder (i.e.,
not adenylated) would also have to be generated in order to analyze profiles composed of non-
adenylated products, which has in fact been accomplished by at least one group (Verheij et al.,
2012). I also wanted to find a polymerase that is compatible with a non-fast instrument
(discussed below). Out of convenience, I preferred to find a hot-start polymerase that was
available in a ready-made master mix, as opposed to one in which the master mix had to be
prepared by adding all components separately (i.e., the polymerase, buffer and dNTPs), but did
not rule those out completely. Hot-start polymerases are also highly recommended to prevent
20
non-specific amplification and stutter (Wang et al., 2009). Cost was also taken into
consideration, as was manufacturer. I was more comfortable evaluating well established
manufacturers with a forensic DNA product line, like QIAGEN and Life Technologies, but I was
also open to evaluating products from other vendors, especially if these products had proven
successful (or at least had been tested and demonstrated potential) in previous studies. For
example, SpeedSTAR™ demonstrated its fast PCR suitability with Profiler Plus and Identifiler
primer sets (Foster & Laurin, 2012; Giese et al., 2009; Laurin & Frégeau, 2012; Vallone et al.,
2008) and is competitively priced compared to similar products from QIAGEN and Life
Technologies. On the other hand, I chose not to evaluate some of the other polymerases noted
in the literature for various reasons, including that they produced blunt end products (i.e.,
Phusion Flash, Phusion and KOD), they were not hot-start polymerases (i.e., Phusion, AB77,
AB95) (Wang, Chang, & Hennessy, 2009), they were not fast polymerases nor were they
buffered for fast PCR (i.e., AmpliTaq Gold) (Applied Biosystems, 2010a), they had been
discontinued (i.e., PyroStart)(Thermo Fisher Scientific, personal communication, August 7,
2013), they exhibited greater cross-reactivity than standard amplification (i.e., AB-1) (Wang et
al., 2009) and/or they generated profiles with additional, uncharacteristic artifacts (i.e., n+1 and
n-2 stutter peaks with AB-3) (Wang et al., 2009).
Therefore, based upon a review of the literature and available products, as well as
recommendations from manufacturers, I selected four reagents to evaluate for fast PCR:
AmpliTaq Gold® Fast PCR Master Mix (AmpliTaq Gold Fast; Life Technologies), KAPA2G™ Fast
Multiplex PCR Kit (KAPA2G; Kapa Biosystems Inc., Woburn, MA), SpeedSTAR and Type-it®
Microsatellite PCR Kit (Type-it; QIAGEN, Valencia, CA). All of these products are either a fast
21
PCR polymerase, include one or are buffered for fast PCR (see Table 1.4), and all but SpeedSTAR
are prepared as a ready-made master mix in which only the primer set and DNA template need
to be added (Applied Biosystems, 2010b; Kapa Biosystems, 2014a; QIAGEN, 2009b; Takara Bio
Inc., 2013). Of these products, only KAPA2G was developed for fast PCR and multiplexing, while
AmpliTaq Gold® Fast and SpeedSTAR were developed for fast PCR, but not multiplexing
(Applied Biosystems, 2010b; Kapa Biosystems, 2014a; QIAGEN, 2009b; Takara Bio Inc., 2013).
Type-it was developed for multiplexing, but not fast PCR, though the polymerase included in
the Type-it Kit (HotStarTaq Plus DNA Polymerase) is a fast polymerase (QIAGEN, 2009b). As an
added curiosity, I was interested in seeing whether fast PCR for forensic primer sets was more
successful based upon whether the product was developed for fast PCR, multiplexing or both.
Table 1.4
Products Selected for Fast PCR Evaluation with the Identifiler Primer Set
Developed for Ready-made Master Mix? Product Polymerase Fast PCR? Multiplexing?
AmpliTaq Gold® Fast PCR Master
Mix
AmpliTaq Gold® DNA Polymerase
Yes No Yes
KAPA2G™ Fast Multiplex PCR Kit
KAPA2G™ Fast HotStart DNA Polymerase
Yes Yes Yes
SpeedSTAR™ HS DNA Polymerase
Same
Yes No No
Type-it® Microsatellite
PCR Kit
HotStarTaq® Plus DNA Polymerase
No Yes Yes
Note. All products assessed include a hot start DNA polymerase that produces adenylated products.
22
Table 1.5
Added Cost per Reaction for Fast Products
Reaction Volume
Product Recommended
Volumea
6µl 5µl 3µl
AmpliTaq Gold® Fast PCR Master Mix
$0.70 $0.21 $0.17 $0.10
KAPA2G™ Fast Multiplex PCR Kit
$0.50 $0.12 $0.10 $0.06
SpeedSTAR™ HS DNA Polymerase
$1.05 $0.13 $0.11 $0.06
Type-it® Microsatellite PCR Kit
$1.12 $0.27 $0.22 $0.13
Note. Added costs are based upon list prices advertised by product manufacturers (Kapa Biosystems, 2014b; Life Technologies, 2014a; QIAGEN, 2014; Takara Bio Inc, 2014). Current first pass processes include 3μl and 6μl reactions, but 5μl reactions are also of interest for second pass (i.e., reamplification). aManufacturer recommended volumes are 20µl, 25µl and 50µl, for AmpliTaq Gold Fast, KAPA2G/Type-it and
SpeedSTAR, respectively.
Table 1.6
Advertised Maximum Ramp Rates (°C/sec) for Various Thermal Cyclers
Thermal Cycler
Block Material
Block Sample
Block Size Heating Cooling Heating Cooling
C1000a 96-well Aluminum 5
b
b b
Mastercyclera 96-well Silver
b 6
b
b b b
Pikoa 24 or 96-well
c Alloy >5 >4.5
b
b
9700d
(9600 Emulation) 96-well
Aluminum Silver Gold
1 b
b
b
9700d
(Std/Max Mode) 96-well
Aluminume
Silverg
Goldg
2.3 2.6 2.6
1.5 1.45
f
1.58h
1.58h
b
Veritid
60-well 96-well
384-well
Aluminum Alloy
Aluminum
3.30 3.90 3.70
b 2.70 3.35 3.10
b
Note. When possible, information was obtained from manufacturers (Applied Biosystems, 2010d; Applied Biosystems, 2010e; Bio-Rad Laboratories, Inc., 2008; Bio-Rad Laboratories, Inc., personal communication, November 4, 2014; Eppendorf, 2005; Life Technologies, personal communication, November 15, 2013; Life Technologies, 2014c; Thermo Fisher Scientific, 2013; Thermo Fisher Scientific, personal communication, November 3, 2014; University of Rhode Island, 2015). Some manufacturers indicated different heating and cooling maximum ramp rates; for those that did not, it is unclear whether the maximum ramp rate is the same for heating and cooling. It should be further noted that additional fast/non-fast instruments and/or other block sizes are available for the majority of the thermal cyclers listed here. aFast thermal cyclers.
bInformation not supplied by manufacturer.
cOnly accepts microplates with a footprint in the
dimension of a microscope slide (for 24- and 96-well blocks). dNon-fast thermal cyclers.
eCalled “Std” mode.
fAs
determined for a 50μl sample. gCalled “Max” mode.
hAs determined for a 10μl sample.
23
With respect to cost, products from well-established, large vendors were more expensive (see
Table 1.5). Cellmark’s current databasing first pass process includes 3µl and 6µl amplification
reaction volumes, thus the added cost per sample is much lower than what it would be for full
reactions, which range from 20µl to 50µl, depending upon the product.
In addition to using fast polymerases, fast thermal cyclers can help reduce PCR
amplification time because they are capable of faster ramp rates than standard thermal cyclers
(see Table 1.6); therefore, not as much time is spent ramping up or down between each PCR
step and cycling can occur more rapidly (Laurin & Frégeau, 2012). As discussed above, fast
thermal cyclers have been used to develop fast PCR processes involving forensic STRs, including
the Piko and C1000 fast thermal cyclers. However, non-fast thermal cyclers should not be ruled
out for use with fast PCR. Non-fast Veriti thermal cyclers have slightly faster ramp rates than
the standard 9700 with an aluminum block, not to mention the fact other sample block types
(e.g., silver and gold-plated silver) are available for 9700s that offer slightly improved ramp
rates over aluminum blocks (see Table 1.6) (Applied Biosystems, 2010d; Applied Biosystems,
2010e; Life Technologies, 2014c). Furthermore, the commonly used 9700 (Butler, 2012, p. 75;
Vallone et al., 2008), is not used to its full potential with regard to ramp rates as a means to
keep PCR amplification parameters the same as that for an older thermal cycler model that had
a significantly lower ramp rate of 1°C/sec (i.e., the GeneAmp® PCR System 9600; 9600; Perkin-
Elmer/Applied Biosystems, Foster City, CA). Manufactures of common forensic STR
amplification kits (e.g., Identfiler, Identifiler Plus and PowerPlex® 16 HS System [PowerPlex 16
HS; Promega]) instruct users to set the 9700 to “9600 Emulation Mode” in order to use the PCR
protocols published in their user manuals for 9700s and the older 9600 (Applied Biosystems,
24
2006a; Applied Biosystems, 2009; Promega, 2014a). For the 9700, Promega also instructs to
reduce ramp rates to 29% for the primer annealing steps and 23% for the extension steps for
PowerPlex 16 HS amplification (Promega, 2014a). Similarly, Applied Biosystems (now Life
Technologies) indicated that Veriti thermal cyclers can be set to either “9700 Max Mode” or
“9600 Emulation Mode” for similar reasons (Applied Biosystems, 2010c). Thus, if the
manufacturer recommended PCR cycling temperatures and times are employed on non-fast
instruments while utilizing their maximum ramp rates, then the PCR time can be reduced by as
much as 63 minutes (or 38%) for PowerPlex 16 HS or 15-16 minutes (9-10%) for Identifiler and
Identifiler Plus, respectively, for 25μl reactions (see Figure 1.2). Additionally, if low volume
amplifications (e.g., 5µl) are used (discussed in more detail below), the same PCR parameters
will take slightly less time compared to that needed for a standard 25µl reaction. However, it
has yet to be determined whether or not using non-fast thermal cyclers in these manners will
generate high quality STR profiles.
Figure 1.2. Reducing amplification time through ramp rates. Total time needed for amplification was determined by entering the manufacturer recommended PCR parameters on the specified thermal cycler using the indicated ramp rates and recording the run time, as determined by the thermal cycler. Substantially more time can be saved by using faster ramp rates for HS because the manufacturer instructs that ramp rates be reduced to 23-29% for various steps of the protocol.
0.00
1.00
2.00
3.00
4.00
5μl 25μl 5μl 25μl 5μl 25μl
ID ID+ HS
Tim
e (
hr)
9600 Mode 9700 Aluminum Block 9600 Mode 9700 Silver Block 9600 Mode 9700 Gold Block
25
However, manufacturers admit that running the same protocol with faster ramp rates
only reduces overall cycling time by minutes (which is consistent with the data in Figure 1.2,
aside from PowerPlex 16 HS), and significantly more time will be saved via a complete
optimization (including selection of reagents, instruments, etc.) that results in the reduction of
the length of time allotted for the various amplification steps (Bio-Rad Laboratories, Inc., 2008).
In practice, this has also been demonstrated, as indicated by the 18 minute fast PCR protocol
developed for the Identifiler primer set with the SpeedSTAR polymerase on the C1000 thermal
cycler (Foster & Laurin, 2012) compared to the 36 minute fast PCR protocol also developed for
Identifiler with a combination of PyroStart and SpeedSTAR polymerases on the 9700 (Vallone et
al., 2008). Thus, even after complete optimization, there was only an 18 minute difference
between fast protocols that employed fast rather than non-fast thermal cyclers.
Despite the added value of using a fast thermal cycler, I chose not to evaluate one for
several reasons. First, thermal cyclers are expensive. For laboratories that run on a tight budget,
it may not be feasible to purchase a fast instrument. Second, it has been demonstrated that fast
PCR is possible with standard, non-fast thermal cyclers (Vallone et al., 2008), and I thought it
would better serve the forensic community to use the 9700 because it is commonly used in
most forensic DNA laboratories. Additionally, I chose to evaluate the non-fast 384-well Veriti
because Cellmark routinely uses it, and it has slightly faster ramp rates than the 9700 (see Table
1.6). Third, as discussed above, using fast thermal cyclers will only reduce amplification time by
mere minutes, and in my option, does not justify the added costs associated with a new
instrument.
26
Thermal cycler ramp rate also plays into a bigger picture for fast PCR – that is, the time
in which it takes to achieve a uniform temperature throughout the sample at each of the
specified temperatures. Because temperature changes are not instantaneous, some lag time
must be built into the amplification parameters to allow for the entire sample to reach the
specified temperature (Butler, 2012, pp. 71-72; Giese et al., 2009). Still, if the time it takes to
reach temperature uniformity can be reduced, then so too can the time allotted for that
particular step. In addition to faster ramp rates contributing towards this goal, utilizing lower
reaction volumes, thin-walled tubes (or plates) and 2-step rather than 3-step PCR will also
contribute. When we apply the first law of thermodynamics, which is an adaption of the law of
conservation of energy, we know that the sample’s temperature change must be the direct
result of the addition (or loss) of heat from an external source (i.e., the thermal cycler and its
block). Moreover, the amount of heat required to impart a temperature change on the sample
can be determined from the well-established equation:
Heat = specific heat x mass x temperature change
For PCR, the heat addition (or loss) will not be instantaneous, but will occur at a certain rate
over a particular amount of time depending upon the quantity of heat required. Additionally,
since mass is directly proportional to volume, we can conclude that less time is required to
impart a temperature change on a smaller volume of a sample as compared to a larger volume
of the same sample, when the other variables remain constant (i.e., amount of heat added or
lost, the sample’s specific heat and the temperature change). This is a generally accepted
concept in molecular biology (Butler, 2012, pp. 71-72). Furthermore, the sample itself is not in
direct contact with the thermal cycler block, but is contained in a tube (or well of a plate), and
27
heat is also needed to impart the same temperature change to the plastic consumable
containing the sample. Thus, the use of thin-walled tubes will further reduce the amount of
time needed for the sample to reach temperature uniformity because such tubes have less
mass than standard tubes and, therefore, less heat (and consequently less time) is needed to
change the temperature of the consumable. Thin-walled tubes are currently the most common
type of tube used for PCR (Butler, 2012, p. 72). Lastly, by combining two of the three steps used
for standard PCR (denaturation, annealing and extension), the number of temperature changes
is reduced by one-third, thereby saving time associated with ramping to a third temperature
and awaiting temperature uniformity of the sample. Use of 2-step PCR cycling combines the
annealing and extension steps into a single annealing-extension step that allows for primer
annealing and extension of PCR products using a single temperature. This approach has been in
practice in the forensic DNA field since the early to mid-2000s, first appearing in real-time
quantitative PCR kits (e.g., the Quantifiler™ Human DNA Quantification Kit, Life Technologies),
followed by STR amplification kits a few years later (e.g., Identifiler Plus) (Applied Biosystems,
2006b; Applied Biosystems, 2009). Thus, these three latter concepts are already established in
the field, but I wanted to note them and attempted to take them a step further for fast PCR.
In practice, these principles are typically not evaluated by how quickly sample
temperature uniformity is achieved, but by evaluating profile quality after reducing the overall
time needed for a fast PCR protocol. For example, most “full volume” reactions for forensic STR
amplifications are 20-50μl (Butler, Advanced Topics in Forensic DNA Typing: Methodology,
2012, p. 72), but reaction volume reduction to 10-13μl has been demonstrated to improve PCR
efficiency and profile quality as compared to the full volume reaction without any modifications
28
to the PCR protocol itself (Gray, Crowle, & Scott, 2014; Leclair et al., 2003). Fast PCR protocols
have also been achieved using 10-15μl reaction volumes, coupled with other modifications
(e.g., fast reagents, fast thermal cyclers, etc.), but resulted in profiles with slightly higher levels
of n-4 stutter (Foster & Laurin, 2012), -A and non-specific products (Vallone et al., 2008) when
compared to those from full reactions. Giese et al. (2009) also reported acceptable levels of n-4
stutter, PHR and -A, but it is unclear how these levels compared to profiles obtained from non-
fast, full volume reactions. With extremely reduced reaction volumes (e.g., 3μl), PCR efficiency
is increased so much using the standard PCR protocol without any modifications (i.e., non-fast
thermal cycler, non-fast reagents), that over-amplification occurs, and as a result, the number
of PCR cycles must be reduced (e.g., from 28 to 26 for Identifiler and Identifiler Plus and from
30 to 28 for PowerPlex 16 HS) in order to obtain profiles with appropriately sized peak heights
and overall quality (A. K. LeFebvre, personal communication, September, 2013).
One last noteworthy factor to consider with regard to fast PCR is that of DNA input.
Reduction in PCR time at any step can make fast PCR less efficient than standard PCR if not
compensated for appropriately with the use of fast reagents, instruments, reduced reaction
volumes, etc. Giese et al. (2009) demonstrated this inefficiency with regard to achieving sample
temperature uniformity imposed by a thermal cycler by developing fast PCR protocols carried
out in a microfluidic biochip versus a 19.6 minute protocol on a fast thermal cycler (the
Mastercycler) and compared both to the 145.1 minute standard PCR protocol. Of note is the
fact that sample temperature uniformity was achieved and maintained for the vast majority of
each of the three PCR steps using standard PCR, but with fast PCR, the sample never reached
any of the desired temperatures (Giese et al., 2009). Consequently, fast tube-based reactions
29
had an approximately 40% decrease in product yield compared to standard PCR (Giese et al.,
2009). Though a similar test of effective sample temperatures could not be obtained using the
microfluidic biochip, samples obtained using the fast biochip protocol had similar product yields
as compared to those obtained using standard PCR (Giese et al., 2009). Thus, it may be
suggested to increase DNA input to further compensate for these inefficiencies, but since -A has
already been noted using independently developed fast PCR procedures (Giese et al., 2009;
Vallone et al., 2008), it seems unwise to add additional DNA template.
Normalized extraction and fast PCR, though somewhat new, are no longer considered
novel ideas, but I ventured into virtually uncharted territory with regard to a detection method
that utilized an unconventional process for forensic STR profiles to further reduce the overall
amount of time need to process DNA samples. Capillary electrophoresis detection of STRs has
been around since the mid-1990s (Wang et al., 1995) and is currently the primary method used
for forensic STR allele separation and detection (Butler, 2012, p. 141). Life Technologies
(formerly Applied Biosystems) is the leading manufacturer of capillary electrophoresis
instruments and has produced various models, including a single capillary ABI Prism 310
Genetic Analyzer, as well as capillary array instruments such as the 3100, 3130, 3700, 3730 and
3500 series Genetic Analyzers, with the number of capillaries ranging from four to 96 per array
(Butler, 2012, pp. 141-142).
Like traditional slab-gel electrophoresis, separation is based on the size and charge of
each molecule and can be controlled by the viscosity of the separation matrix, distance traveled
through the matrix, time of travel, voltage and current (Buel, LaFountain, Schwartz, &
Walkinshaw, 2001; Butler, 2012; Moretti et al., 2001). Instead of using agarose or
30
polyacrylamide as the separating matrix, the capillaries use a liquid polymer. Life Technologies
manufactures three such polymers (POP-4® Polymer [POP-4], POP-6™ Polymer [POP-6] and
POP-7™ Polymer [POP-7]) and four array lengths (22cm, 36cm, 50cm and 80cm) for use with
their instruments (Applied Biosystems, 2007). POP-4 and POP-6 are less viscous than POP-7,
and each polymer is designed for a different purpose (Life Technologies, personal
communications, November 11, 2014 and December 5, 2014; Life Technologies, 2014d). Life
Technologies recommends using POP-4 for DNA sequencing and fragment analysis (i.e.,
microsatellites and SNPs) for fragments <500bp, and POP-4 is commonly used for human
identification (HID; e.g., forensic applications) purposes. POP-6 is recommended for standard
and rapid DNA sequencing, but is currently the only option for long sequence fragments
(>500bp) on 310 and 3100 Genetic Analyzers (Life Technologies, 2014e). Like POP-4, POP-7 is
recommended for DNA sequencing and fragment analysis, but offers the added capability of
sequencing up to ~1000bp. However, it is not compatible with the 310 or 3100 instruments
(Life Technologies, 2014f). The shortest array (22cm) is used for fragment analysis and the
longest (80cm) is used for sequencing, whereas the 36cm and 50cm arrays can be used for
either (Applied Biosystems, 2007). Table 1.7 displays polymer and array length combinations for
various types of separation and detection applications, as recommended by Life Technologies.
It appears that most laboratories in the forensic DNA community follow Life
Technologies’ recommendation to utilize POP-4 with a 36cm array (on various models of
Genetic Analyzers) for forensic STR/human identification (HID) applications (Foster & Laurin,
2012; Laurin & Frégeau, 2012; Myers et al., 2012; Verheij et al., 2012), but some are also using
or have used POP-6 with a 36cm array (Vallone et al., 2008).
31
Table 1.7
Recommended Polymer Type and Array Length Combinations
Type of Run Polymer Array Length (cm) Approximate Run Time
(min)
Ultra Rapid Sequencing POP-4 POP-7
36 40 35
Rapid Sequencing POP-6 POP-7
36 60
Fast Sequencing POP-7 50 60
Standard Sequencing POP-4 POP-6 POP-7
50 100 150 120
Long Read Sequencing POP-4 POP-7
80 210 170
High Throughput, Small Size Fragment Analysis
POP-4 22 20
Standard Fragment Analysis
POP-4 POP-7 POP-4 POP-6 POP-7
36 36 50 50 50
45 35 65 90 50
Note. Adapted from the 3130/3130xl Getting Started Guide (Applied Biosystems, 2007, pp. 132-133).
Others using POP-6 appear to use a 50cm array based upon a run time of 2.75 hours,
though the array length is not indicated (Koumi et al., 2004). POP-6 offers increased resolution
as compared to POP-4 due to its higher viscosity, and though not necessary, it could be
beneficial because using the HID conditions recommended by Life Technologies (POP-4 with a
36cm array), complete separation (Buel et al., 2001) of alleles differing in size by one base-pair
cannot be achieved (see Figures 1.3 and D.15). Using POP-4 with a 36cm array, Dr. Aaron
LeFebvre of Cellmark’s databasing facility also indicated that on a rare occasion, he has
observed extremely poor resolution of larger amplicons differing by one base-pair (e.g., alleles
13.4 and 14 at the Penta-D locus, which are ~420bp), such that the analysis software could not
differentiate between the two alleles (personal communication, November 12, 2014).
32
Figure 1.3. One base-pair resolution. This representative Identifiler allelic ladder was obtained from a 3130xl Genetic Analyzer using traditional POP-4/36cm detection and demonstrates that complete 1bp resolution cannot be achieved. Specifically, at the TH01 locus, alleles 9.3 and 10 are not completely resolved (see circled alleles).
Thus, it is imperative that any changes that are implemented in order to decrease the detection
time do not reduce resolution as compared to that which is currently achieved via POP-4 on a
36cm array.
To date, no studies involving significant reductions in capillary electrophoresis run time
have been identified, but Dr. LeFebvre said he had conducted some preliminary work involving
POP-6 on a 22cm array in 2008 that yielded results that looked promising for forensic STR usage
(personal communication, December, 2012). Furthermore, Azco Biotech manufactures polymer
alternatives to Life Technologies’ POP-4, POP-6 and POP-7, called NanoPOP4, NanoPOP6 and
NanoPOP7, respectively (Azco Biotech, Inc., Oceanside, CA) (Azco Biotech, Inc., 2012). The
33
NanoPOP polymers include a higher percentage of a key proprietary ingredient found in Life
Technologies’ polymers, which results in faster run times than can be achieved using Life
Technologies’ polymer with as good or better resolution (A Perlman, personal communication,
January 26, 2013). Other polymer alternatives have been evaluated, such as MCLAB’s
NanoPOP-4™ and were deemed suitable for DNA analysis with regard to precision and
resolution, but these products were not developed to have decreased run times (Sciarretta,
Berlin, Smith, & Dawson Cruz, 2011). Thus, I decided to evaluate decreased capillary
electrophoresis run times on a 3130xl using POP-6 on a 22cm array, NanoPOP4 on a 36cm
array, NanoPOP6 on a 22cm array and NanoPOP6 on a 36cm array.
Table 1.8
Parameters for the Default HID Run Module
Defined in Data Collections v2.0
Parameter Name Value Acceptable Range
Oven Temp. Oven_Temperature 60 18-65°C Poly Fill Vol. Poly_Fill_Vol 6500 6500-38000steps
Current Stability Current_Stability 5 0-2000μAmps PreRun Voltage PreRun_Voltage 15 0-15kVolts
PreRun Time Pre_Run_Time 180 1-1000sec Injection Voltage Injection_Voltage 3 1-15kVolts
Injection Time Injection_Time 10 1-600sec Voltage # of Steps Voltage_Number_Of_Steps 40 1-100nk Voltage Step Int. Voltage_Step_Interval 15 1-60sec Data Delay Time Data_Delay_Time 1 1-3600sec
Run Voltage Run_Voltage 15 0-15kVolts Run Time Run_Time 1500 300-14000sec
Note. The default run module for HID fragment analysis using POP-4 on a 36cm array is the “HIDFragmentAnalysis36_POP4_1” run module on the 3130xl (Applied Biosystems, 2007; Life Technologies, personal communication, June 11, 2014).
34
Alon Perlman of Azco Biotech indicated that their NanoPOP polymers were originally
developed for use with a 310 Genetic Analyzer (personal communication, January 26, 2013).
Therefore, it was anticipated that some of the run module parameters may need to be adjusted
when these polymers are used on a 3130xl. Additionally, Life Technologies does not supply a
run module for POP-6 on a 22cm array, so those parameters will also need to be determined.
Default run module parameters for HID fragment analysis using POP-4 on a 36cm array
(recommended by Life Technologies for HID usage) is provided in Table 1.8. As with traditional
slab-gel electrophoresis voltage, current, run time and temperature are all related to each
other in capillary electrophoresis. Ohm’s law (V = I x R; voltage is equal to the product of
current and resistance) demonstrates the relationship between voltage and current, as they
relate to resistance, and it is the current that dictates the speed at which DNA molecules travel
through the matrix (Butler, 2012, p. 149). Temperature also has a significant impact on
resistance because the two are inversely proportional to one another (Butler, 2012, p. 149);
thus temperature also impacts current and, therefore, time. For example, if a higher oven
temperature is utilized during the electrophoresis process, then the viscosity of the polymer will
be reduced, thereby reducing the resistance. Since current is held constant (i.e., defined by the
run module), it will take less time for a fragment to reach the detection cell in this environment
with a higher temperature/lower resistance as compared to environments with a lower
temperature/higher resistance. On the other hand, resolution improves when fragments are
allowed to separate at a slower pace over a longer distance (Butler, 2012, p. 151; Moretti et al.,
2001). Thus, using higher voltage, current and/or temperature to decrease run time, will in turn
decrease resolution. What remains to be seen is whether decreasing run time, coupled with the
35
use of a polymer that is more viscous (i.e., POP-6) than the polymer that is normally used for
forensic STR detection (i.e., POP-4), can result in acceptable levels of resolution for HID
applications.
Switching to a more viscous polymer and a shorter array may also influence peak height.
As a sample travels through the polymer, its individual alleles separate based upon size (Butler,
2012, p. 143). Multiple copies of the same allele migrate together, but as the distance traveled
increases, these alleles begin to separate slightly, as is indicative of the sample appearing as a
two-dimensional “peak” comprised of a height and width. Buel et al. (2001) observed a
broadening in peak width (measured at half height) with a change from POP-4 to POP-6
polymer when processed on a 36cm array. Likewise, increasing the array length from 36cm to
50cm also provided an increase in peak width using POP-4 (Buel et al., 2001). Given that the
quantity of an allele can be calculated by the peak’s area, it is inferred that if the same
quantities are processed using different methods, their peak areas should be the same, and
when peak width increases (as described above), then peak height should decrease.
Furthermore, based upon the observations above, it can be inferred that changing from a 36cm
array to a 22cm array should impart a reduction in peak width and, therefore, an increase in
peak height when the same type of polymer is used. However, when this array change is
coupled with a change from POP-4 to POP-6 polymer, the magnitude of this increase in peak
height is uncertain because changing from POP-4 to POP-6 appears to reduce peak height based
on Buel et al.’s observation that peak width increases. Depending upon the degree with which
each of these factors (i.e., a change in array length and polymer type) effect peak width and,
therefore peak height, it is possible that processing samples with POP-6 on a 22cm array may
36
result in either an increase or a decrease in peak height as compared to processing with POP-4
on a 36cm array. Thus, at this point, actual impact on peak height remains to be determined. If
there is a significant difference in peak height between the traditional POP-4/36cm array
detection method and a proposed, shorter detection method, then peak heights can be
adjusted accordingly by modifying injection parameters. Specifically, injection voltage and time
can be altered to control the quantity of sample that is electrokinetically injected (Butler, 2012,
p. 145), which in turn will impact peak height. Moretti et al. (2001) demonstrated that increases
in peak height are proportional to increases in injection time, but warns that resolution is
decreased via peak broadening with prolonged injections (15-20sec). Furthermore, injection
times less than five seconds are generally not utilized for HID applications (Foster & Laurin,
2012; Moretti et al., 2001). Likewise, injection voltage is directly proportional to peak height
(Butler, 2012, p. 145) and may be increased or decreased accordingly, but is generally 2kV or
3kV for HID applications (personal observation).
In summary, there is an expressed interest and need to improve sample processing
efficiency via a reduction in processing time and cost for forensic DNA samples. I proposed a
three step overhaul of an existing process that was utilized by an established DNA crime
laboratory that:
1. Eliminated the need for quantification and pre-amplification dilution of all samples by
modifying the currently used ChargeSwitch extraction procedure, such that DNA
concentrations were normalized within a narrow amplification target range during the
extraction and purification process,
37
2. Reduced amplification time by developing fast PCR protocols for three commonly used
forensic STR kits (Identifiler, Identifiler Plus and PowerPlex 16 HS),
3. Reduced STR profile detection time by developing a quicker detection method using
unconventional polymer/array length combinations, while still maintaining a suitable
degree of allele resolution.
With these changes, not only was processing time reduced, but cost savings manifested in the
form of increased sample throughput per hourly labor and a decrease in reagent and supply
costs, as well as instrument usage. Specifically, with regard to reagent and supply costs, no cost
increase was associated with utilization of a normalized extraction. It should be noted that the
small per sample reagent cost associated with purchasing a supplemental fast PCR polymerase
was more than offset by the decrease in cost associated with eliminating the quantification and
dilution steps, as well as eliminating the need to purchase supplemental AmpliTaq Gold (which
is needed for standard PCR using the Identifiler kit) and extending the life of the other two
amplification kits tested (Identifiler Plus and PowerPlex 16 HS)1. For the new detection method
selected, cost savings was achieved through the use of less polymer per detection run. It should
be noted that in order for this cost savings to manifest in a working laboratory, the entire bottle
of polymer must be used prior to expiring once it has been placed on the instrument (typically
one week), otherwise the degree of cost savings is unclear.
1 The master mixes included in these two kits were the limiting reagents when following the manufacturer
recommended PCR protocols, thereby unused primer sets were discarded. However, using a supplemental fast PCR polymerase eliminated the need for all other kit components except for the primer sets, thereby allowing more amplification reactions per kit because the entire tube of primer set could be used.
38
Overall, the these improvements will benefit the forensic DNA community, as well as
the numerous other molecular biology fields that routinely analyze STR profiles via PCR
amplification and capillary electrophoresis detection.
39
CHAPTER 2
MATERIALS AND METHODS
2.1 Project Overview
This research project improved processing efficiency for forensic DNA samples via a
reduction in processing time and costs. Robust processes were selected from a private forensic
DNA crime laboratory for such improvements, specifically for “first pass” high throughput
processing of buccal swabs and Buccal DNA Collectors™ in their databasing facility. This project
consisted of three primary components, undertaken in the following order: development of a
quicker capillary electrophoresis detection method using a 3130xl Genetic Analyzer;
development of fast PCR protocols using the Identifiler, Identifiler Plus and PowerPlex 16 HS
primer sets; and extraction normalization using the ChargeSwitch® Forensic DNA Purification
Kit. A quicker detection procedure was developed first because some preliminary work had
been performed by Dr. Aaron LeFebvre in 2008 and would result in a substantial reduction in
processing time with little to no cost increase. Fast PCR protocols needed to be developed prior
to normalized extractions for two reasons. First, I needed to assess STR profile quality obtained
from fast amplification of current, non-normalized extractions. Second, I needed to determine
optimal DNA input range and volume utilized in the fast amplification reaction. Only after both
of these were completed, could I then develop a normalized extraction procedure that would
target the appropriate range of input DNA (in the allotted volume) required for fast PCR.
When applicable, 95% confidence intervals (CI) for averages are given in tables or
displayed as error bars in graphical figures. All tests for significance utilized α=0.05.
40
2.2 Development of a Quicker Capillary Electrophoresis Detection Method for Forensic STR Profiles
2.2.1 Initial Evaluation
Three polymers and two array lengths were selected for an initial evaluation: NanoPOP4
(Azco Biotech, Inc.) on a 36cm array (Life Technologies); NanoPOP6 (Azco Biotech, Inc.) on a
22cm (Life Technologies) and a 36cm array; and POP-6 (Life Technologies) on a 22cm array.
Three lots of NanoPOP4 and NanoPOP6 were obtained from Azco Biotech, Inc. to test lot-to-lot
polymer quality. Spectral and regular run modules were primarily based on default HID
POP-4/36cm array run modules, with influences from default POP-4/22cm array and
POP-6/36cm array run modules (no default run modules exist for POP-6/22cm array) (Applied
Biosystems, 2007), and were minimally modified to assess individual run module settings (see
Table 1.8), including a necessary reduction in polymer array fill from 6500 to 5900 steps for
22cm arrays only. See Appendix A (Tables A.1-A.4) for specific run module parameters tested.
Spectral calibration matrix standards for dye sets F (4-dye Life Technologies
amplification kits, such as COfiler and Profiler Plus), G5 (5-dye Life Technologies amplification
kits, such as Identifiler and Identifiler Plus) and D (arbitrarily named for this dissertation; 4-dye
Promega amplification kits, such as PowerPlex® 16 System [PowerPlex 16; Promega] and
PowerPlex 16 HS) were prepared as described by the manufacturer. Standards were then
processed on a 3130xl using traditional (POP-4/36cm) and modified spectral calibration run
modules (see Table A.1) in order to obtain spectral run modules resulting in acceptable spectral
calibrations for the various tested polymer/array length platforms.
Diluted amplification product (1μl) obtained from COfiler, Identifiler, Identifiler Plus,
Profiler Plus, PowerPlex 16 or PowerPlex 16 HS amplifications were combined with 10μl of a
41
formamide:size standard mixture (see Table 2.1), denatured for three minutes at 95°C, and
subjected to a three minute snap freeze. Prepared samples were then processed on a 3130xl
pre-heated to the specified run temperature (55°C, 60°C or 63°C) using a traditional HID
POP-4/36cm run module and an experimental polymer/array/run module combination (see
Table A.2). Spectral calibrations were either set to a passing spectral obtained using
POP-4/36cm array or to a passing spectral obtained from an experimental polymer/array/run
module combination. Resulting profiles were analyzed using GeneMapper® ID v3.2 with
analysis thresholds of 75-150rfu and assessed for overall quality and total run time. Resolution
was qualitatively assessed through visual signs of peak broadening (Butler et al., 2004). ILS
sizing quality is a unitless value (0-1.0, 1.0 being perfect) calculated by the GeneMapper® ID
software and is used to describe the migration quality of a sample based on the migration of
known-sized fragments of the size standard. Thus, it too was assessed. The polymer/array
combination(s) exhibiting the most promise for a reduced detection time, while maintaining
high profile quality, was selected for further optimization.
Table 2.1
Preparation of Amplification Product for Capillary Electrophoresis Detection
Product Reaction Volume Water Added Size Standard Hi-Di™ Formamide
a:
Size Standard (μl)
COfiler Profiler Plus
3μl 4μl GeneScan™ 500 ROX™
a
9.7:0.3
Identifiler Identifiler Plus
3μl 4μl GeneScan™ 500 LIZ™
a
9.8:0.2
PowerPlex 16 PowerPlex 16 HS
3μl 4μl ILS 600b 9.5:0.5
Note. Amplifications were performed on a Veriti thermal cycler. All amplification product was diluted with the volume of water indicated above, and 1μl of diluted product was added to 10μl of the specified formamide:size standard mixture. aManufactured by Life Technologies.
bManufactured by Promega.
42
2.2.2 Optimization of an Alternative Detection Method Using POP-6 (22cm Array)
Alternative detection methods involving either NanoPOP4 or NanoPOP6 were
abandoned based upon the performance of NanoPOP4 during the initial evaluation. POP-6/
22cm run modules with two of the three initially assessed oven temperatures (60°C and 63°C),
and various other modifications (see modules 13, 14 and 21 in Table A.5), showed potential as
successful detection methods, but each needed adjustments to improve STR profile quality.
Additional refinements discussed here focused on injection parameters to obtain desired peak
heights and reduce the frequency of called “pull-up” peaks, as well as improvements to ILS
sizing quality; resolution was again qualitatively monitored through visual signs of peak
broadening.
2.2.2.1 Modifications to Injection Voltage and Time
Modifications to POP-6/22cm run modules were performed in an effort to obtain full
profiles with adequate peak heights, free of oversaturation and with minimal detectable pull-up
(in frequency of occurrence and percent pull-up [ 20%] compared to the source peak).
Identifiler Plus amplification product was prepared for detection as discussed previously (see
Table 2.1). A full plate (89 samples, two positive controls, two negative controls and three allelic
ladders) was processed using traditional POP-4/36cm detection using an established protocol
from Cellmark (see G5 in Table A.3) with a 3kV 7sec injection, followed by two injections worth
of samples (29-31 samples and one allelic ladder) from the same plate using POP-6/22cm
detection with five experimental run modules. All experimental modules were based on one
that was initially evaluated with a 60°C oven temperature (see module 13 in Table A.5) and
43
varied by injection voltage (2kV and 3kV) and injection time (4-20sec)(see Table A.6). All
resulting POP-4 profiles were analyzed with GeneMapper® ID using a 150rfu threshold to
establish the STR profile. POP-6 profiles were then analyzed (150rfu threshold) and were first
reviewed for the occurrence of oversaturated peaks and allelic dropout. If any profiles exhibited
oversaturation and/or allelic dropout, samples from that run module were not analyzed
further. If full profiles were obtained without oversaturation from all samples for a particular
run module, then that data set was further assessed for the occurrence of called pull-up peaks
(excluding those that overlapped with other alleles or n-4 stutter), as well as peak height, and
compared to corresponding results from the POP-4 data set.
2.2.2.2 Modifications to Run Voltage, Run Temperature and Current Stability
Additional modifications to POP-6/22cm run modules were performed in an effort to
obtain profiles with ILS sizing qualities comparable to that of traditional POP-4/36cm detection.
Identifiler Plus amplification product was prepared for detection as discussed previously (see
Table 2.1). Two injections worth of samples (30 samples and two allelic ladders) were processed
using traditional POP-4/36cm detection using an established protocol from Cellmark (see G5 in
Table A.3) with a 3kV 7sec injection. From these two injections, either the first or second
injection was reprocessed using POP-6/22cm detection with eleven experimental run modules.
All experimental modules were based on the one that was selected from the previous section
with a 2kV 7sec injection (see module 5 in Table A.6) and varied by run voltage (12kV, 13kV or
14kV), run temperature (60°C or 63°C) and current stability (4µA, 5µA or 6μA). Data delay time
and run time were also modified, as necessary, in response to changes in speed of separation
44
dictated by changes in other parameters (see Table A.7). All resulting profiles were analyzed
with GeneMapper® ID using a 150rfu threshold and were examined for ILS sizing quality; total
run time was also noted for each run module.
Next, a total of 12 Identifiler Plus allelic ladders were injected (across three injections)
using seven different POP-6/22cm run modules with varying run voltages (12-15kV) and oven
temperatures (60°C or 63°C) based on those processed above (see Table A.8). Four ladders per
injection were positioned in 12 different wells (see Figure 2.1) so as to account for sizing
variability within an injection, between injections and between different capillaries. Data was
also collected from 12 Identifiler Plus ladders processed using traditional POP-4/36cm
detection. All resulting ladders were analyzed with GeneMapper® ID using a 150rfu threshold
and were assessed for ILS sizing quality and precision of base-pair sizing for all alleles; total run
time was all noted for each run module.
2.2.3 Validation of an Alternative Detection Method Using POP-6 (22cm Array)
Next, I validated the POP-6/22cm array capillary electrophoresis detection method (see
Table A.9) for use with Life Technologies 5-dye and Promega 4-dye amplification kits.
Figure 2.1. Positioning of allelic ladders. Allelic ladders (blue) were loaded in a total of 12 of the 16 capillaries and processed using seven POP-6/22cm run modules.
1 2 3 4 5 6 7 8 9 10 11 12
A
B
C
D
E
F
G
H
45
Validation consisted of a comparison to traditional POP-4/36cm detection (including
reproducibility/concordance of allele calls, precision and resolution), as well as a sensitivity
study to obtain a usable range of injection times. Despite some initial testing with amplification
product from Life Technologies’ 4-dye kits, this validation did not include run modules for them
because they are an older technology and are becoming obsolete.
All detection plates were prepared as described previously. Samples were analyzed
using GeneMapper® ID with 150rfu (PowerPlex 16) or 180rfu (Identifiler Plus) thresholds unless
otherwise noted and were given a pass or fail rating using highly desired profile characteristics
regarding detection issues (see Table 2.2).
Table 2.2
Pass/Fail Detection Guidelines
Criteria Pass Fail
% Alleles Detecteda 100% <100% Maximum Peak Height 6000rfu >6000rfu Oversaturation None Any occurrence PHR at all heterozygous loci 50% <50% at any locus
Pull-up 20% >20%
Stutter (n-4, n+4) 20% >20% -A None Any occurrence Elevated Baseline None Any occurrence
Migration All allele calls correct Any occurrence of poor migrationb
Injection Failure None Any occurrence Loss of Resolutionc None Any occurrence
Spikes 1 occurrencef >1 occurrencee Note. Some of these restrictions are more related to amplification than detection (e.g., stutter). “Occurrence” relates to a detected peak that is called using GeneMapper® ID. aAt the defined threshold.
bThat results in off-ladder (OL) or miscalled allele(s).
cPoor ILS or unresolved peaks.
dOkay
if in multiple dye channels but occur at the same base-pair size. eAt different base-pair sizes.
46
2.2.3.1 Comparison to Traditional POP-4/36cm Detection
Two full plates (each consisting of 89 samples, one extraction blank, two positive
amplification controls, one negative amplification control and three allelic ladders) of both
Identifiler Plus (3µl reaction volume) and PowerPlex 16 (3µl reaction volume) samples that had
been previously processed using traditional POP-4 on a 36cm array (3kV 7sec) were
reprocessed with POP-6 on a 22cm array. It should be noted that Identifiler Plus detection
plates were prepared from amplification product that was approximately one month old
(stored at -20°C), whereas fresh, unstored PowerPlex 16 amplification product was used.
Following GeneMapper® ID analysis, all alleles were assessed for concordance to demonstrate
reproducibility of allele calls using POP-6 detection.
Resolution was quantitatively assessed for each detection method’s ability to resolve
alleles differing by one base-pair within all allelic ladders (n=6 per amplification type) and
samples (n=1 for Identifiler Plus and n=2 for PowerPlex 16) exhibiting such alleles. Resolution
was calculated using the valley value (Vv) approach described by Buel et al. (2001) in which the
height of the valley (point where the two peaks merge) is divided by the peak height of the
taller allele. This method was preferred over traditional resolution calculations that require the
peak width (at half height) described by Buel et al. (2001) or Heller (1999) because peak width
cannot always be determined for alleles differing by one base-pair. Furthermore, since this
assessment was limited to a very small sample size, resolving power, or resolution length (RSL;
Heller, 1999), was also calculated for allele peaks in the Identifiler Plus and PowerPlex 16 allelic
ladders according to the formula:
RSL = Wh/(dX/dM) ≈ Wh/(∆X/∆M)
47
in which Wh is the width at half height (measured in time), ∆X is the difference between peaks
(measured in time) and ∆M is the size difference between alleles (measured in bp). RSL was
calculated for all allele peaks, except 9.3 and 10 at TH01 because width at half height could not
always be determined due to overlap of those peaks. Assessing resolution in this manner
identifies the smallest difference between allele sizes (bp) that can be resolved. Ideally, this
value should be 1, indicating that the detection method can resolve alleles differing by one
base-pair.
Lastly, ILS sizing quality, peak heights and pass rates were compared between data sets.
2.2.3.2 Precision
A total of 11 Identifiler Plus and 11 PowerPlex 16 HS allelic ladders were used to assess
allele sizing and precision with POP-6 on a 22cm array. Three or four ladders per injection were
positioned in 11 different wells (see Figure 2.2) so as to account for intra-injection, inter-
injection and inter-capillary sizing. All resulting allele peaks were analyzed with GeneMapper®
ID using a 150rfu threshold.
A) Identifiler Plus B) PowerPlex 16 HS
Figure 2.2. Positioning of allelic ladders for validation. Allelic ladders (blue) were loaded in a total of 11 of the 16 capillaries.
1 2 3 4 5 6 7 8 9 10 11 12
A
B
C
D
E
F
G
H
1 2 3 4 5 6 7 8 9 10 11 12
A
B
C
D
E
F
G
H
48
2.2.3.3 Sensitivity and Usable Injection Times
From the comparison study, one injection eqivalent of samples exhibiting dropout using
the 2kV 7sec injection (wells A09-H10 from the second Identifiler Plus and the first PowerPlex
16 detection plates, each including 13-14 samples and an allelic ladder) were reinjected using
2kV 4sec and 2kV 15sec injections with POP-6. Following analysis, sensitivity of the detection
methods was assessed via profile completeness, peak height and oversaturation; pass rates
were also determined. These data were compared to that collected using a 2kV 7sec injection
during the POP-4/POP-6 comparison study.
2.3 Development of Fast PCR Protocols for Identifiler, Identifiler Plus and PowerPlex 16 HS
2.3.1 Initial Evaluation
Four amplification reagents/kits (AmpliTaq Gold® Fast PCR Master Mix, KAPA2G™ Fast
Multiplex PCR Kit, SpeedSTAR™ HS DNA Polymerase and Type-it® Microsatellite PCR Kit) were
evaluated for fast PCR. A buccal swab cutting of known origin was extracted using the EZ1® DNA
Investigator Kit (QIAGEN) on a BioRobot® EZ1 instrument and quantified using the Quantifiler®
Human DNA Quantification Kit (Life Technologies) according to manufacturer protocols, except
that a half reaction was used for quantification. This sample was then amplified in triplicate
(~1ng DNA) using the above four reagents in 3μl amplifications on a 384-well Veriti thermal
cycler, coupled with fast PCR amplification parameters (see Tables 2.3 and 2.4).
49
Table 2.3
Fast PCR Reaction Composition for Initial Evaluation
AmpliTaq Gold Fast KAPA2G SpeedSTAR Type-it 1.5µl AmpliTaq Gold® Fast PCR
Master Mix (2X) 0.65µl Identifiler Primers 0.85µl DNA Template + Water
1.5µl KAPA2G™ Fast Multiplex Mix
0.65µl Identifiler Primers 0.85µl DNA Template + Water
0.3µl 10X Fast Buffer I 0.24µl dNTPs 0.015µl SpeedSTAR™ HS 0.65µl Identifiler Primers 1.80µl DNA Template + Water
1.5µl Type-it Multiplex PCR Master Mix
0.65µl Identifiler Primers 0.85µl DNA Template + Water
Note. Reaction composition was based on established 3μl Identifiler amplifications (using standard PCR) at Cellmark, as well as manufacturer recommendations.
Table 2.4
Fast PCR Thermal Cycling Parameters for Initial Evaluation
PCR Step AmpliTaq Gold Fasta KAPA2G
b SpeedSTAR
c Type-it
d
Polymerase Activation 95°C 10min 95°C 3min 95°C 1min 95°C 5min 26 Cycles of: Denaturation Annealing Extension
96°C 5sec
59°C 10sec 68°C 10sec
95°C 15sec 59°C 30sec 72°C 30sec
98°C 5sec
59°C 15sec 72°C 10sec
96°C 5sec 59°C 5sec
68°C 15sec Final Extension 72°C 1min 72°C 1min 72°C 1min 72°C 1min Hold 4°C 25°C 25°C 25°C Total Time 35min 50min 29min 30min
Note. Thermal cycling parameters were based on established 3μl Identifiler amplifications (using standard PCR) at Cellmark, as well as manufacturer recommendations. aManufacturer recommendations for fast PCR, but not specific for multiplexing, were modified by slightly
increasing denaturation, annealing and extension times from 3sec, 3sec and 5sec, respectively. bManufacturer
recommendations for fast PCR and multiplexing. cManufacturer recommendations for fast PCR, but not specific for
multiplexing. dManufacturer recommendations for HotStarTaq® Plus DNA Polymerase as part of Fast Cycling PCR
Kit, not the Type-it Microsatellite PCR Kit, which was developed for multiplexing, but not fast PCR.
Following amplification, amplification product was diluted with 4µl water. From the
resulting 7µl product dilution, 1µl was combined with Hi-Di™ Formamide (Life Technologies)
and GeneScan™ 500 LIZ™ (Life Technologies) size standard (10µl of a 10:0.1 mix), and
subsequently detected via capillary electrophoresis on a 3130xl Genetic Analyzer using a 3kV
7sec injection (POP-4, 36cm array). STR profiles were analyzed using GeneMapper® ID with a
75rfu threshold and were assessed for amplification quality, including profile completeness,
allele concordance, allele peak height, intra- and inter-locus balance, -A, stutter and baseline
50
noise. An STR profile previously obtained from the known sample (processed with standard,
non-fast PCR using Identifiler Plus) was used as a comparison for allele concordance and overall
quality to the profiles obtained using fast PCR.
It should be noted that the sample that was used for this initial evaluation was extracted
and quantified with different chemistries than were used for the process that was ultimately
selected for improvement at Cellmark in their databasing unit. Therefore, all following
evaluations utilized ChargeSwith extracts quantified with PicoGreen to remain consistent with
the databasing process at Cellmark.
2.3.2 Development of Four Fast PCR Protocols for Identifiler
Based on the initial evaluation of the four fast PCR protocols, KAPA2G™ Fast Multiplex
PCR Kit produced the best STR profiles with the Identifiler primer set and appeared to require
the least amount of modifications to achieve optimization. Therefore, a fast PCR protocol was
developed for it first, and information obtained from that was used to streamline development
of fast PCR protocols for AmpliTaq Gold Fast, SpeedSTAR and Type-it. Following development of
all four protocols, a comparison study was performed to determine which protocol would be
selected for use with other STR primer sets. For all studies, ChargeSwitch extracts from buccal
swabs or buccal collector punches were amplified using various amounts of input DNA, as
determined by PicoGreen quantification. Promega 9947A (0.2ng/μl, 0.3ng/μl or 0.4ng/μl) was
utilized as a positive amplification control, whereas water was used for a negative amplification
control for all amplifications. Amplifications were carried out in a total volume of 3μl on a Veriti
thermal cycler. Fast PCR reaction composition and thermal cycling parameters remained
51
unchanged for the specified product under evaluation (as given in Tables 2.3 and 2.4), with
specific modifications to the thermal cycling parameters noted below. Detection and analysis
were performed as described in the above section; additional analyses are also discussed in
detail in the 2.3.2.5 Comparison of Four Fast PCR Protocols for Identifiler section. During
development of each of the four protocols, I aimed for full profiles free of oversaturation, -A
and non-specific amplification (NSA) products, average peak heights of 750-1500rfu, average
PHR >85%, minimal occurrences of PHR <50% (preferably no occurrences) and inter-locus
balance of ~0.35 (as measured via the coefficient of variation of locus peak height to the
profile’s total sum of peak height ratios). Additionally, percent stutter and percent pull-up had
to be below 20%; actual number of occurrences of stutter, pull-up and elevated baseline were
qualitatively assessed for fast PCR protocol development, but were quantitatively assessed
during protocol comparison.
2.3.2.1 Identifiler Fast PCR with the KAPA2G™ Fast Multiplex PCR Kit
For all studies, anywhere from one (in triplicate) to 24 samples were amplified using
~0.25-1.25ng DNA (in ~0.25ng increments), with appropriate controls. In an effort to eliminate -
A, I evaluated final extension times of 1min, 2min, 3min, 4min, 5min and 10min. Primer
specificity was evaluated through the use of 59°C, 61°C and 63°C annealing temperatures.
Various means of reducing total amplification time were investigated through the use of 2-step
PCR cycling (combined annealing-extension steps of 30sec, 35sec and 40sec), shorter
denaturation time (5sec, 10sec and 15sec) and shorter activation time (1min, 2min and 3min).
See Table B.2 for specific amplification parameters tested.
52
2.3.2.2 Identifiler Fast PCR with the AmpliTaq Gold® Fast PCR Master Mix
For all studies, two samples were amplified using ~0.25ng, 0.50ng and 0.75ng DNA, each
in triplicate (unless otherwise indicated), with appropriate controls. It should be noted that
amplification reaction composition was modified slightly to maintain the same primer
concentration (20%; reduced from 22% in the initial evaluation) that is used for standard
Identifiler amplification (see Table 2.5). To eliminate -A, final extension times of 1min, 5min,
10min and 13min were evaluated. Primer specificity was evaluated through the use of 59°C,
61°C and 63°C annealing temperatures. Improvements to amplification efficiency (i.e.,
increased peak height, reduced allelic dropout, improve inter-locus balance and elimination of
PHR <50%) were evaluated by using 2-step PCR cycling (combined annealing-extension steps of
40sec, 50sec, 1min and 2min), 3-step PCR cycling (15sec, 30sec, 45sec and 1min annealing;
30sec, 45sec and 1min extension) and longer denaturation time (5sec, 10sec, 15sec, 20sec and
1min). Due to some unexpected difficulties, this product was also evaluated using standard
Identifiler amplification thermal cycling parameters (see Table 2.6). In an effort to reduce total
amplification time, I also evaluated shorter activation times (1min, 3min and 10min). A 25°C
final hold was also evaluated in comparison to the 4°C hold initially evaluated (using ~0.75ng
DNA in triplicate from two samples). See Table B.3 for specific amplification parameters tested.
2.3.2.3 Identifiler Fast PCR with the SpeedSTAR™ HS DNA Polymerase
For all studies, two samples were amplified using ~0.25ng, 0.50ng and 0.75ng DNA, each
in triplicate (unless otherwise indicated), with appropriate controls. It should be noted that
amplification reaction composition was modified slightly to maintain the same primer
53
concentration (20%; reduced from 22% in the initial evaluation) that is used for standard
Identifiler amplification (see Table 2.5). Primer specificity was evaluated through the use of
59°C, 61°C and 63°C annealing temperatures. Improvements to amplification efficiency (i.e.,
improved inter-locus balance and elimination of PHR <50%) were evaluated by using 2-step PCR
cycling (5 and 10sec denaturation; 25sec, 30sec, 40sec, 50sec and 1min combined annealing-
extension) versus 3-step PCR cycling (15sec, 20sec and 25sec annealing; 10sec and 20sec
extension). Later indications of -A were assessed through longer final extension times (1min,
10min and 13min). See Table B.4 for specific amplification parameters tested.
2.3.2.4 Identifiler Fast PCR with the Type-it® Microsatellite PCR Kit
For all studies, two samples were amplified using ~0.25ng, 0.50ng and 0.75ng DNA, each
in triplicate (unless otherwise indicated), along with controls. It should be noted that
amplification reaction composition was modified slightly to maintain the same primer
concentration (20%; reduced from 22% in the initial evaluation) that is used for standard
Identifiler amplification (see Table 2.5). Improvements to amplification efficiency (i.e., obtain
complete, balanced profiles) were evaluated by using longer times for each of the PCR steps
compared to those initially evaluated, including 2-step PCR cycling (combined annealing-
extension steps of 55sec, 1min 5sec and 1min 15sec,) versus 3-step PCR cycling (45sec,
annealing; 30sec extension) and denaturation times of 10sec, 20sec and 30sec. In an effort to
reduce total amplification time, I also evaluated shorter activation times (1min, 2min and
5min). See Table B.5 for specific amplification parameters tested.
54
2.3.2.5 Comparison of Four Fast PCR Protocols for Identifiler
Twenty-five buccal swabs from different donors were extracted using the ChargeSwitch
kit (quarter reaction), quantified using PicoGreen and normalized via dilution with water to
0.667ng/μl. Two of these samples were also serially diluted (ten dilutions) to determine the
optimal DNA input range. One sample was diluted such that 5.86x10-3ng to 3.00ng of DNA was
amplified, while the other was diluted such that 1.95x10-3ng to 1.00ng of DNA was amplified. All
normalized samples and serial dilutions (in triplicate) were amplified using each of the four fast
PCR protocols that were developed for Identifiler, as well as using standard PCR with the
standard Identifiler reagents. All amplifications utilized 0.9μl DNA in a 3μl total reaction volume,
for a final DNA input of ~0.6ng (aside from the serial dilutions); some fast PCR protocols
required the addition of water to each sample to bring those reactions up to volume (see Table
2.5). All samples and reagents were dispensed into a 384-well amplification plate using the
Nanodrop™ Express. The corresponding thermal cycling parameters for each of the five
amplification data sets (four fast and one standard) were employed for amplification on a 384-
well Veriti thermal cycler (see Table 2.6). Profile detection and analysis were performed as
discussed previously. The four fast PCR protocols were compared to each other and to standard
Identifiler amplification via their optimal DNA input ranges (determined by sensitivity,
reproducibility, inter-locus balance, intra-locus balance, stutter, pull-up, -A, non-specific
amplification and baseline noise), stochastic threshold, precision of allele sizing, stutter and an
optimization check consisting of 25 samples. Each of these analyses is discussed in more detail
below.
55
Table 2.5
Final Fast PCR Reaction Compositions Compared to Standard PCR
AmpliTaq Gold Fast
KAPA2G SpeedSTAR Type-it Standard Identifiler
1.5µl AmpliTaq Gold® Fast PCR Master Mix (2X)
0.6µl Identifiler Primers 0.9µl DNA Template
1.5µl KAPA2G™ Fast Multiplex Mix
0.6µl Identifiler Primers 0.9µl DNA Template
0.3µl 10X Fast Buffer I 0.24µl dNTPs 0.015µl SpeedSTAR™ HS 0.6µl Identifiler Primers 0.9µl DNA Template 0.945µl Water
1.5µl Type-it Multiplex PCR Master Mix
0.6µl Identifiler Primers 0.9µl DNA Template
1.15µl PCR Reaction Mix 0.05µl AmpliTaq Gold 0.6µl Identifiler Primers 0.9µl DNA Template 0.3µl Water
Note. Reaction composition was based on established Identifiler 3μl amplifications (using standard PCR) at Cellmark, as well as manufacturer recommendations, including a 20% final concentration of the Identifiler primer set.
Table 2.6
Final Fast PCR Thermal Cycling Parameters Compared to Standard PCR
PCR Step AmpliTaq Gold Fast
KAPA2G SpeedSTAR Type-it Standard Identifiler
Polymerase Activation
95°C 10min 95°C 1min 95°C 1min 95°C 5min 95°C 11min
26 Cycles of: Denaturation Annealing Extension
96°C 10sec 61°C 45sec 68°C 45sec
95°C 5sec
61°C 40sec
98°C 5sec
61°C 25sec 72°C 20sec
96°C 30sec
59°C 1min 15sec
94°C 1min 59°C 1min 72°C 1min
Final Extension 72°C 13min 72°C 10min 72°C 13min 72°C 10min 60°C 60min Hold 25°C 25°C 25°C 25°C 4°C
Total Time 1hr 19min 43min 49min 1hr 14min 2hr 42min Note. Final thermal cycling parameters for each of the four fast PCR protocols for Identifiler are compared to the standard parameters for 3μl amplifications (established at Cellmark).
2.3.2.5.1 Determination of the Optimal DNA Input Ranges
2.3.2.5.1.1 Sensitivity
Determination of each amplification’s optimal DNA input range was multifaceted and
was based upon both serial dilutions that were amplified in triplicate. Sensitivity range was a
key component; this was defined as the lowest DNA input amount from which full profiles were
56
obtained from the majority of the samples, through the highest DNA input from which full
profiles were obtained from the majority of the samples, using a 75rfu threshold. Full profiles
had to be obtained at each of the DNA inputs throughout the entire sensitivity range. Unless
otherwise noted, all additional analyses for the determination of the optimal DNA input range
only utilized data from within the sensitivity range.
2.3.2.5.1.2 Reproducibility
Reproducibility of peak height was assessed using triplicate amplifications at each DNA
input amount via calculating the average, standard deviation and coefficient of variation (CV)
for each allele’s peak height. Coefficient of variation for the allele peak heights was then
averaged per DNA input amount. Average CVs of 0.350 were indicative of the desired level of
reproducibility.
2.3.2.5.1.3 Inter-locus Peak Balance
Inter-locus balance underwent a two-fold assessment2. First, I attempted to quantify the
degree of preferential amplification based upon locus size, specifically looking for profiles in
which the smallest-sized amplicons (Amelogenin) had total peak heights approximately four or
more times greater in magnitude than the largest-sized amplicons (e.g., CSF, D18 or D2), but
this proved to be quite challenging. Two groups have used a coefficient of preferential
amplification/hybridization (CPA) to calculate amplification bias with respect to determining
2 It should be noted that this was the only time inter-locus peak balance underwent a two-fold
assessment. All other references to inter-locus peak balance evaluation refer only to general inter-locus balance via CV of LPH:TPH values and not imbalance due to preferential amplification based on locus size using CPALS.
57
allele frequencies from pooled DNA (Hoogendoorn et al., 2000; Yang, Pan, Lu, & Fann, 2005;
Yang et al., 2006); I adapted their approach for use with the Identifiler primer set in an attempt
to calculate a coefficient of preferential amplification based on locus size (CPALS; see Appendix
B). For each profile obtained from the various DNA inputs, I plotted the sum of each locus’ peak
height (locus peak height, or LPH) against locus (arranged in order of increasing size). Linear
regression lines were generated for each profile plot; I assumed that if the R2 value was
moderate to strong (>0.3), then the degree of inter-locus balance due to preferential
amplification (i.e., the CPALS) could be quantified by dividing the negative slope by the
y-intercept. Based on models I developed, I determined that CPALS values of >0.050 suggested
that the locus peak height for the smallest-sized locus (Amelogenin) was approximately four or
more times greater than that one of the largest-sized locus (CSF) and was therefore deemed
unacceptable.
Next, overall locus-to-locus balance, regardless of locus size, was assessed by calculating
the sum of each locus’ peak height divided by the profile’s total sum of peak heights to obtain
each locus’ locus:profile peak height ratio (or locus peak height to total sum of peak heights,
LPH:TPH). Then, the coefficient of variation for each profile’s LPH:TPH was calculated and
averaged for each template amount. Although there are no guidelines regarding acceptable
levels of LPH:TPH CV values, I sought averages of 0.350 during the development of each fast
PCR method based on inter-locus balance obtained using standard Identifiler amplifications (3μl
reaction volume).
58
2.3.2.5.1.4 Intra-locus Peak Balance
Intra-locus balance was assessed by calculating peak height ratios at heterozygous loci;
these were averaged at each of the DNA input amounts, and additionally, average instances of
PHR <50% per profile were calculated. It should be noted that these fast PCR procedures will be
used for single source, reference samples, therefore PHR tolerances of 50% are acceptable,
compared to the traditional 60% that is typically utilized for casework samples to help identify
mixtures (FBI, 2011).
2.3.2.5.1.5 Other Amplification Artifacts
Stutter (n-4, n+4, n-8), pull-up, -A, non-specific amplification (NSA) and baseline were
also assessed as part of the DNA input range determination. Only the peaks that were actually
called using GeneMapper® ID were included in these analyses; locus specific stutter filters as
defined within the software were used, but no other filters were applied. For each stutter, pull-
up or -A peak detected, its peak height was divided by the peak height of the true allele from
which it originated in order to calculate the percentage of that peak (defined as “percent
stutter,” “percent pull-up” or “percent –A,” respectively). These values were averaged for each
template amount assessed. Additionally, average instances of stutter, pull-up, -A, NSA and
baseline peaks per profile were also calculated. At times it was difficult to differentiate between
low-level NSA and elevated baseline, given that all profiles assessed originated from only two
individuals.
59
An optimal DNA input range was selected using the above criteria. The optimal range
did not have to exhibit perfection for all criteria, but needed to exhibit near-acceptable to
acceptable levels for all, with sensitivity and PHR carrying the most weight.
2.3.2.5.2 Stochastic Threshold
The stochastic assessment utilized the lowest DNA input amount from the sensitivity
range down to the highest DNA input amount in which complete dropout consistently occurred.
All profiles were reanalyzed using the laboratory’s limit of detection (25rfu) and examined for
extreme dropout probability (Gill, Puch-Solis, & Curran, 2009) by counting the number of
dropout instances where the surviving allele of a heterozygote locus is higher than the limit of
detection. Then, using various potential stochastic thresholds, the number of dropout
occurrences was calculated and graphed against each potential threshold to determine an
acceptable stochastic threshold in which zero occurrences of dropout occurred.
2.3.2.5.3 Precision
Precision of allele sizing was assessed via 9947A amplification controls loaded in
triplicate within a single injection on a detection plate (capillaries 6, 9 and 14), along with an
Identifiler allelic ladder. This set was injected three times to assess intra- and inter-injection
precision. Base pair sizes for each allele were averaged for each triplicate within an injection
(intra-injection assessment) and also for each triplicate within a single capillary (inter-injection
assessment). Standard deviation (<0.15) and differences in allele sizing (±0.5bp) were used to
assess precision according to manufacturer recommendations.
60
2.3.2.5.4 Stutter Assessment
In addition to the limited stutter assessment that was performed on called stutter peaks
from the sensitivity range data, a more in depth assessment of stutter was performed on the
optimal DNA input range, the 25 optimization check samples and all 9947A positive
amplification controls. For this analysis, stutter filters were not applied during GeneMapper® ID
analysis, thereby allowing the detection of all peaks (including stutter) that occurred using a
25rfu threshold. Average percent stutter was calculated for n-4, n+4 and n-8 stutter for each
locus and amplification method. Any stutter peak that overlapped with a true allele was
excluded from this analysis, as were overlapping n-4 and n+4 peaks. If an n±4 peak from the
first allele overlapped with an n-8 peak from the second allele, is was averaged with n±4 stutter
only, not n-8. Stutter peaks coinciding with pull-up from another dye channel were also
excluded.
2.3.2.5.5 Optimization Check
Profiles obtained from 25 different samples served as a small optimization check to
assist with selecting the best fast PCR protocol. Since development of two of the fast PCR
protocols (AmpliTaq Gold Fast and SpeedSTAR) experienced significant problems obtaining PHR
consistently above 50% and there was some evidence that PHR <50% were associated with
large amplicon-sized loci and/or large allele separation (i.e., most notably for AmpliTaq Gold
Fast and the sample exhibiting the largest allele difference [24bp], which happened to be at a
large amplicon locus [D18]), I attempted to select individuals with varying degrees of allele
separation (differing by up to nine alleles, or 36bp) at D18 and other large sized loci (e.g., D2
61
and FGA). Thus, in addition to assessing PHR as discussed above for the optimal DNA input
range determination, it was also assessed for signs of greater intra-locus imbalance due to the
magnitude of allele separation and/or locus itself. Additionally, these profiles were assessed for
allele concordance, profile completeness (percent full profiles and average percent alleles
detected per sample), inter-locus balance (as discussed above) and first pass success rates. First
pass success rate was defined as the percentage of passing profiles obtained during the first
round of testing (i.e., without re-extraction, re-amplification, re-injection, etc.). Passing profiles
must have all alleles detected at or above threshold, all PHR 50%, no called stutter peaks
>20%, no called pull-up peaks >20% and no -A; see Table 2.7 for a complete list of
passing/failing guidelines.
2.3.3 Development of Additional Fast PCR Protocols using KAPA2G™ Fast Multiplex PCR Kit
Based on the performance of the four fast PCR protocols as compared to that of
standard Identifiler amplification, KAPA2G was selected for the development of additional fast
PCR protocols, including 5µl Identifiler, 6µl Identifiler, 3µl Identifiler Plus and 3µl PowerPlex 16
HS amplifications. The same general criteria with regard to profile quality that were used during
the development of the 3µl Identifiler fast PCR protocols were also used for these protocols.
2.3.3.1 Development of 5µl and 6µl Identifiler Fast PCR Protocols
For all studies, one or two samples were amplified using ~0.25ng, 0.50ng, 0.75ng and
1.0ng DNA, each in triplicate (unless otherwise noted), with appropriate controls. Amplification
reaction composition consisted of 2.5µl KAPA2G™ Fast Multiplex Mix, 1.0µl Identifiler primer
62
set and 1.5µl DNA template. Amplification was carried out in a 96-well plate in a 9700, and
thermal cycling parameters initially began with the same parameters used for the 3µl fast PCR
reaction developed for KAPA2G (except with a 5min final extension). Improvements to
amplification efficiency were evaluated by using the “Max” versus “9600 Emulation” modes on
the thermal cycler, assessing different numbers of amplification cycles (27 and 28, compared to
26 used for the 3µl reaction), assessing longer annealing/extension times (45sec and 50sec,
compared to 40sec) and assessing a longer final extension (10min compared to 5min; n=24).
See Table B.6 for specific amplification parameters tested.
Table 2.7
First Pass Analysis Guidelines
Criteria Pass Fail
% Alleles Detecteda 100%, no signs of mixture <100% PHR at all heterozygous loci 50% <50% at any locus
Pull-up 20% >20%
Stutter (n-4, n+4) 20% >20%
Stutter (n-8) 2 occurrences, 5% >2 occurrences ( 5%) or any occurrence >5%
Trialleles and microvariantsb None Any occurrence -A None Any occurrence +A None Any occurrence Elevated Baseline Occurrences at 3 loci Occurrences at >3 loci Non-specific Amplification None Any occurrence
Oversaturation 2 oversaturated peaks >2 oversaturated peaks
Migration All allele calls correct Any occurrence of poor migrationc
Injection Failure None Any occurrence Loss of Resolutiond None Any occurrence Spikes 1 occurrencee >1 occurrencef
Note. “Occurrence” relates to a detected peak that is called using GeneMapper® ID. aAt the defined threshold.
bMust be reprocessed to confirm; however, given that all samples that were tested
either originated from a known source or were also processed using a current procedure for comparison purposes, these would only fail if non-concordant with the known profile or profile obtained from the current process.
cThat
results in OL or miscalled allele(s). dPoor ILS or unresolved peaks.
eOkay if in multiple dye channels but occur at the
same base-pair size. fAt different base-pair sizes.
63
Development of the 6µl Identifiler fast PCR protocol was based upon the 5µl protocol
and required changes to reaction composition only, which consisted of 3.0µl KAPA2G™ Fast
Multiplex Mix, 1.2µl Identifiler primer set and 1.8µl DNA template.
Lastly, 24 samples were processed using the 5µl and 6µl Identifiler fast amplification
protocols, and fast profiles were compared to those obtained using current Identifiler standard
amplification protocols used at Cellmark (see Table B.1).
2.3.3.2 Development of 3µl Identifiler Plus Fast PCR Protocol
For all studies, two samples were amplified using ~0.25ng, 0.50ng and 0.75ng DNA, each
in triplicate (unless otherwise noted), along with controls. Amplification reaction composition
consisted of 1.5µl KAPA2G™ Fast Multiplex Mix, 0.6µl Identifiler Plus primer set and 0.9µl DNA
template. Amplification was carried out on a 384-well plate in a Veriti thermal cycler. The initial
thermal cycling parameters that were utilized were based upon the protocol that was
developed for Identifiler fast PCR using KAPA2G. Improvements to amplification efficiency
began with assessing final extension times (1min, 5min and 10min) to prevent the formation of
-A peaks and different annealing/extension temperatures (59°C, 61°C and 63°C) to prevent non-
specific amplification, followed by efforts to minimize amplification time via reducing activation
length (3min, 2min and 1min), denaturation (15sec, 10sec and 5sec) and annealing/extension
(60sec and 50sec). Eighty-eight samples were processed using two potential 3µl Identifiler Plus
fast amplification protocols in order to identify which performed better, and these fast profiles
were compared to those obtained using the current 3µl Identifiler Plus standard amplification
protocol used at Cellmark (see Tables B.1 and B.7).
64
2.3.3.3 Development of 3µl PowerPlex 16 HS Fast PCR Protocol
For all studies, two samples were amplified using ~0.25ng, 0.50ng and 0.75ng DNA, each
in triplicate (unless otherwise noted), with appropriate controls. Amplification reaction
composition consisted of 1.5µl KAPA2G™ Fast Multiplex Mix, 0.3µl PowerPlex 16 HS primer set
and 1.2µl DNA template/water. Amplification was carried out in a 384-well plate in a Veriti
thermal cycler. The initial thermal cycling parameters that were utilized were based upon the
3µl protocol that is currently used for PowerPlex 16 HS at Cellmark and information learned
from development of the other fast PCR protocols using KAPA2G. Improvements to
amplification efficiency began with assessing final extension times (1min, 5min and 10min) to
prevent the formation of -A peaks, 2-step PCR cycling with different annealing/extension
temperatures (58°C, 60°C and 62°C) to prevent non-specific amplification and increasing ramp
rates to 100% for annealing and extension steps. These were followed by efforts to minimize
amplification time via reducing annealing time (30sec and 15sec), extension (30sec, 20sec and
15sec), denaturation (15sec, 10sec and 5sec), initial activation (2min and 1min) and final hold
(4°C and 25°C. Eighty-nine samples were processed using this 3µl PowerPlex 16 HS fast
amplification protocol, and fast profiles were compared to those obtained using the current 3µl
PowerPlex 16 HS standard amplification protocol used at Cellmark (see Tables B.1 and B.8).
2.3.4 Validation of Fast PCR Protocols using KAPA2G™ Fast Multiplex PCR Kit
I next validated each of the five fast PCR protocols that I developed using KAPA2G (3µl,
5µl and 6µl Identifiler, 3µl Identifiler Plus and 3µl PowerPlex 16 HS; see Table B.9) for use with
buccal cell samples (swabs and Buccal DNA Collector™ punches). Validation parameters
65
included sensitivity, reproducibility, inter- and intra-locus peak balance, stutter, pull-up, -A and
baseline studies to determine the optimal range of input DNA, as well as, stochastic, precision,
stutter, automation (large sample sets), contamination, lot-to-lot variation and storage
condition assessment. Studies comparing each fast PCR method to its respective standard
amplification were performed during the development process for each method (see above),
and therefore were not repeated during the validation.
2.3.4.1 Determination of the Optimal Range of Input DNA
Two known, highly heterozygous male samples were manually serially diluted such that
the total amount of input DNA (ng) in each amplification ranged from approximately
5.86x10-3ng to 3.00ng; each dilution was amplified in triplicate. As described previously in
section 2.3.2.5 Comparison of Four Fast PCR Protocols for Identifiler, optimal DNA input ranges
were determined for each of the five fast PCR protocols based upon sensitivity, reproducibility,
inter- and intra-locus peak balance and presence of artifacts called by the analysis software.
2.3.4.2 Stochastic Threshold, Precision, Stutter, Automation and Contamination
Stochastic, precision and stutter (n±4 and n-8) studies were also performed as described
previously, with the following three modifications to the precision study. First, precision was
assessed for each of the primer sets using the 3µl fast PCR protocols (additional precision
studies for 5µl and 6µl Identifiler fast PCR were deemed unnecessarily redundant). Second, the
capillaries 3, 6 and 16 were used for Identifiler/Identifiler Plus, while 4, 7 and 14 were used for
PowerPlex 16 HS. Third, fast results were not compared to standard PCR, but were instead held
66
to the ±0.5bp sizing difference and <0.15 standard deviation requirement for alleles up to
250bp, as advertised by Life Technologies to obtain correct allele calling using POP-4/36cm
detection and GeneMapper® ID analysis software. Next, for each primer set, a large number
(86-89) of samples (previously amplified using standard PCR) were amplified using fast PCR; STR
profiles were assessed for concordance, profile completeness, peak height, intra- and inter-
locus peak balance and first pass success rate. Contamination studies examined all negative
amplification controls for signs of contamination (25rfu threshold).
2.3.4.3 Lot-to-Lot Variation
Lot-to-lot variation of KAPA2G™ Fast Multiplex PCR Kit was assessed using five known
samples and controls via amplification with 3µl Identifiler fast PCR reactions (~0.75ng template
DNA) using three different lots of KAPA2G™ Fast Multiplex Mix. All three lots were stored
frozen until use and were all used prior to their expiration dates. The preparation dates of the
lots were not known, but Lot 1 had the earliest expiration date, followed by Lot 2 and then Lot
3. Following GeneMapper® ID analysis, STR profiles were assessed for profile completeness,
peak height, intra- and inter-locus peak balance.
2.3.4.4 Storage Conditions
KapaBiosystems indicates that KAPA2G™ Fast Multiplex Mix may be stored at 4°C for up
to one month and is stable for up to fifty freeze/thaw cycles. To assess reagent performance as
a result of number of thaws and storage at 4°C, I thawed a total of 12 aliquots 1-3 times and
used them either immediately or following storage at 4°C for one week, two weeks or one
67
month to amplify five known samples and controls with 3µl Identifiler fast PCR reactions
(~0.75ng template DNA). Following GeneMapper® ID analysis, STR profiles were assessed for
profile completeness, peak heights and intra- and inter-locus peak balance.
2.3.5 Post-Validation Modifications to Identifiler Fast PCR
Additional testing of the 3µl and 6µl Identifiler fast PCR protocols continued during the
extraction normalization portion of the project, and signs of low-level, non-specific
amplification (separate from the low-level Amelogenin artifact identified during validation) and
-A (6µl reaction only) were noted, such that annealing temperature and final extension time
were revisited for these two fast protocols.
For 3µl Identifiler fast amplifications, 86 ChargeSwitch extracts (buccal swab cuttings),
an extraction blank and positive/negative amplification controls were amplified using the
validated amplification protocol, except with the use of a 63°C annealing/extension
temperature rather than the validated 61°C. All resulting profiles were examined for profile
quality with respect to completeness, peak height and intra-/inter-locus peak balance.
For 6µl amplifications, final extensions of 13min, 15min, 20min and 23min were
assessed in comparison to 10min. From those data sets, the 20min final extension was selected
and used in combination with a 63°C annealing/extension step to amplify 12 ChargeSwitch
extracts (Buccal DNA Collector™ punches), an extraction blank and positive/negative
amplification controls; no other modifications were made to the 6µl amplification protocol.
68
2.4 Development of a Normalized Extraction using the ChargeSwitch® Forensic DNA Purification Kit
Numerous experiments were conducted in order to modify the ChargeSwitch extraction
procedure utilized in the Cellmark databasing laboratory, such that DNA concentrations from
buccal swab cuttings (swabs) and Buccal DNA Collector™ punches (punches) would consistently
permit fast amplification using a specified target range of DNA (0.375-1.50ng – as determined
by the fast PCR validations discussed above – unless otherwise noted) and result in a high first
pass success rate. In this section, it should be assumed that normalized extraction was used
unless otherwise specified. Buccal samples were collected from known, high and low shedders.
During the normalized extraction procedure, bead volumes were substantially reduced on a per
sample basis; in order to maintain pipetting accuracy, a mixture with sufficient beads and
Purification Buffer (both from the ChargeSwitch kit) for the entire extraction batch was
prepared, and 100µl of the mixture was dispensed into each sample well (compared to the
current process of adding beads and 100µl Purification Buffer separately). Quantification via
PicoGreen was always performed during development in order to assess the progress of
extraction normalization, as well as during validation, but is no longer a required step since the
new first pass procedure has been validated. Many of the early developmental experiments
processed samples using fast and standard PCR (without dilution) in order to determine if
profile defects were the result of the normalized extraction versus fast PCR; early studies
tended to use 3µl amplifications (HS/ID+/ID) for swabs and 6µl amplifications (ID only) for
punches to mirror the Cellmark processes, but 3µl amplifications were ultimately tested and
validated for both sample types and all three primer sets. It should be noted that although 5µl
Identifiler fast amplifications were validated, these were not further tested with the first pass
69
process. Traditional POP-4/36cm array (POP-4) and alternative POP-6/22cm array (POP-6)
detection were often used as well throughout testing. Any references to the “current” first pass
procedure refer to non-normalized extraction followed by quantification and dilution, then
standard PCR amplification and POP-4 detection. Noteworthy experiments undertaken are
discussed in more detail below.
2.4.1 Initial Evaluation
One buccal swab was cut into nine pieces (see Figure 2.3) and extracted using the
ChargeSwitch® Forensic DNA Purification Kit using a range of bead and elution volumes. The
four top quarter cuttings were used in combination with bead volumes of 5.00µl, 1.00µl, 0.50µl
and 0.25µl, along with a 60µl elution volume. The four side quarter cuttings were used in
combination with the same bead volumes, but with an 80µl elution volume. The single bottom
cutting was extracted with a 2µl bead volume and an 80µl elution. DNA concentrations were
reviewed in order to determine which had concentrations within the target range for 3µl
amplifications; these were amplified using PowerPlex 16 HS, Identifiler Plus and Identifiler fast
and standard 3µl amplifications and detected using POP-4.
Figure 2.3. Location of swab cuttings.
Top of swab – cut into quarters
Side of swab – cut into quarters
Bottom of swab – used as one sample
70
Though direct comparisons of DNA concentrations could not be made between the
various extraction methods (because different elution volumes were used) using different
portions of the swab, total yields were be compared. This design was a quick assessment of
cellular distribution on the swab and conclusions drawn from data mining regarding bead
volume, elution volume, predicted DNA concentrations and STR profile quality using fast and
standard PCR without post-extraction normalization (i.e., quantification and dilution).
2.4.2 Optimization of Normalized Extraction for Buccal Swab Cuttings and Buccal DNA Collector™ Punches
The initial evaluation of a normalized extraction procedure showed a great deal of
potential for success. Further optimization studies (see Table C.1) included several small test
batches (samples from as few as four individuals processed in triplicate or from 10-20
individuals processed in duplicate) to a few large test batches (samples from 35-42 individuals,
Either processing one or two cuttings/punches each) using buccal swabs or punches (see Figure
2.4 for sample collection). Volumes of Proteinase K (5.0µl or 10µl), beads (0.25-5.0µl), Wash
Buffer (2 x 125µl, 250µl/125µl or 2 x 250µl), Elution Buffer (60µl, 80µl or 120µl) and DNA used
in the amplification reaction (0.9µl for 3µl fast; 0.9µl or 1.2µl for 3µl standard; 0.5-1.8µl for 6µl
fast; 0.7-2.4µl for 6µl standard), as well as incubation length (1.5hr or overnight) and including
or excluding a 5sec post-incubation vortex, were all tested with buccal swabs and/or punches.
As studies progressed, the same DNA volume was used for fast and standard amplification for
purposes of side-by-side comparisons within the same study. Several injection parameters were
also used for detection: POP-4, 3kV 4-12sec for 3µl amplifications; POP-6, 2kV 4-12sec for 3µl
amplifications; POP-4, 2kV 5-12sec and 3kV 7-10sec for 6µl amplifications; and POP-6, 2kV
71
5-12sec for 6µl amplifications. Profiles were analyzed with a 75rfu threshold and evaluated for
quality as discussed previously (i.e., profile completeness, concordance, peak height, inter- and
inter-locus peak balance, artifacts and first pass success rate).
A) Buccal DNA Collector™ Punches
B) Buccal Swab Cuttings
Figure 2.4. Buccal punch (6mm) and swab cutting locations (optimization). Three punches were only collected from buccal collectors during the first optimization study; all following studies utilized one (locations 1 or 2) or two (locations 1 and 2) punches only. 2.4.3 Validation of Normalized ChargeSwitch Extractions Within the New First Pass Processes
Normalized ChargeSwitch extraction procedures (see Table C.2) were validated for
buccal swabs and Buccal DNA Collector™ punches (6mm) as part of the first pass procedure
(see Figure 2.5 for collection locations), including fast PCR using 3μl (Identifiler, Identifiler Plus
and PowerPlex 16 HS; 0.9µl DNA amplified) and 6μl (Identifiler; 1µl DNA amplified; punches
only) amplifications and detection on a 3130xl Genetic Analyzer with the traditional
POP-4/36cm array platform (3kV 7sec for swabs and 2kV 7sec for punches) or an alternative
POP-6/22cm array platform (2kV 7sec for both sample types). Using the normalized extraction
1 3 2
2 1
3
72
process, first pass procedures will allow quantification and dilution to be eliminated once this
process is implemented in a databasing setting. However, quantification via PicoGreen was
performed during the validation for information purposes only; no dilutions were performed
for normalized extraction samples. Validation consisted of sensitivity, reproducibility,
comparison to non-normalized extraction, automation (large sample sets) and contamination
studies.
A) Buccal DNA Collector™ Punches
B) Buccal Swab Cuttings
Figure 2.5. Buccal punch (6mm) and swab cutting locations (validation). Normalized extractions utilized punches from location 1 or 2, while current non-normalized extractions obtained punches from any of the four locations. Both normalized and current non-normalized extractions utilized cuttings from any of the four locations (position 4 located on the back side of the side, similar in a position as 3).
2.4.3.1 Sensitivity
A highly controlled sensitivity study for the normalized extraction could not be designed
2 1
4 3
1 3 2 4
73
for multiple reasons. First, the normalized extraction was designed to forgo dilution; therefore I
did not want to dilute samples in order to obtain desired input DNA amounts in case dilution
itself diluted out inhibitors that may be present from the extraction. Second, I did not want to
attempt to apply known amounts of DNA onto swabs or buccal collectors, because doing so
would require the use of pre-washed/pre-treated cells that are free of other naturally occurring
oral substances that would normally be deposited on the samples during collection and could
skew results by making an extraction appear to be more efficient. I wanted to gather sensitivity
information from a process that was identical in practice to the normalized extraction
procedure I developed. Therefore, I collected buccal samples from a variety of individuals (low,
moderate and high shedders), totaling 170 swab and 68 punch samples, and processed them
using the normalized extraction procedure and PicoGreen quantification method. All of these
were processed with fast PCR (3µl PowerPlex 16 HS, Identifiler and Identifiler Plus for swabs
and punches; 6µl Identifiler for punches only) and POP-6 detection, while 86 of the swabs and
all of the punches were also processed using POP-4 detection. The ability to obtain a full profile
was compared to DNA concentration to determine the sensitivity of each fast PCR procedure
using DNA extracts obtained from the normalized extraction procedure.
2.4.3.2 Reproducibility
Eighty-three buccal swabs and 35 punches were processed in duplicate using the
normalized ChargeSwitch extraction procedures; 43 swabs and all 35 of the punches were
processed in duplicate within an extraction, while the remaining 40 swabs were processed in
duplicate between two separate extractions two weeks apart. Samples were quantified using
74
PicoGreen. DNA concentrations were compared between sample duplicates and among all of
the same sample type to assess reproducibility of DNA recoveries using the normalized
extractions.
2.4.3.3 Comparisons to Current Process
Comparisons between the normalized and current non-normalized extraction procedure
were performed using high quality swabs, old degraded swabs and buccal collector punches.
2.4.3.3.1 High Quality Swabs
Duplicate cuttings (n=86 per set) were obtained from high quality buccal swabs that had
been collected from 43 individuals within four months prior to testing. One set was extracted
using the normalized ChargeSwitch extraction for swabs, followed by fast and standard 3μl
amplifications using PowerPlex 16 HS, Identifiler Plus and Identifiler and POP-4 or POP-6
detection. A second set was processed using the current first pass process for each individual
amplification kit.
2.4.3.3.2 Old, Degraded Swabs
Duplicate cuttings (n=22 per set) were obtained from old, degraded buccal swabs that
had been collected from 22 individuals approximately four years prior to testing. One set was
extracted using the normalized ChargeSwitch extraction for swabs, followed by 3μl Identifiler
Plus fast amplifications and POP-6 detection (2kV 10sec). The other set was processed using the
current first pass process for 3µl Identifiler Plus amplification.
75
2.4.3.3.3 Buccal DNA Collector™ Punches
Duplicate punches (n=35 per set) were obtained from Buccal DNA Collectors™ that had
been collected from 35 individuals within two months prior to testing. One set was extracted
using the normalized ChargeSwitch extraction for punches, followed by 3μl and 6µl fast
amplifications using PowerPlex 16 HS, Identifiler Plus and Identifiler and POP-4 or POP-6
detection. The remaining set was processed using the current first pass process for each
individual amplification kit (3µl for PowerPlex 16 HS, Identifiler Plus and Identifiler; 6µl for
Identifiler).
2.4.3.3.4 Analysis
Bead binding capacities of the normalized and current, non-normalized extraction
procedures were compared for each data set. For high quality swabs, four first pass processes
were compared:
1. Normalized extraction, fast PCR, POP-4 detection
2. Normalized extraction, fast PCR, POP-6 detection
3. Normalized extraction, standard PCR, POP-4 detection
4. Current extraction, quantification, dilution, standard PCR, POP-4 detection
Old, degraded swabs were processed with procedures 2 and 4 only, while punches were
processed with procedures 1, 2 and 4. The first two processes were the new first pass options
utilizing normalized extraction, the third was an intermediate between current and new first
pass options that highlights the impact of the normalized extraction and the last was the
76
current first pass process. When applicable, each of these processes were grouped by
amplification kit for the three sample types (the degraded samples were only amplified with
Identifiler Plus), thereby generating twelve data sets for high quality swabs and punches, but
only two data sets for old, degraded swabs. Pass/fail rates were compared from the data sets
per sample type using a 75rfu threshold and first pass analysis guidelines (modified to
accommodate the low-level artifact identified at Amelogenin during Identifiler fast PCR
validation). Following analysis, profiles were assessed for concordance, profile completeness,
peak heights, intra- and inter- peak balance and first pass success rates.
2.4.3.4 Contamination Assessment
Both of the normalized extraction procedures utilized different bead volumes (0.5µl and
1.0µl) and final elution volumes (60µl and 80µl). Therefore, both protocols were assessed for
contamination. For each, buccal samples (42) and reagent blanks (48) were arranged in a
checkerboard pattern on a 96 well plate (the additional six empty wells were place holders for
amplification controls and allelic ladders required for downstream processing) and were
extracted as described previously. The first contamination plate (0.5µl beads and 60µl elution
volume) was amplified using 3µl fast PCR reactions with all three primer sets, while the second
contamination plate (1.0µl beads and 80µl elution volume) was amplified using 3µl Identifiler
Plus fast PCR reactions; both were detected via POP-6. All samples and blanks were evaluated
for signs of contamination (a 25rfu threshold was applied to reagent blank wells).
77
CHAPTER 3
RESULTS AND DISCUSSION
3.1 Development of a Quicker Capillary Electrophoresis Detection Method for Forensic STR Profiles
3.1.1 Initial Evaluation
3.1.1.1 NanoPOP4 (36cm Array)
Detection using NanoPOP4 on a 36cm array resulted in repeated run failures for the first
and third polymer lots tested due to current levels above the maximum limit, preventing
electrophoresis altogether. Electrophoresis was able to complete using the second lot tested,
which allowed me to evaluate the product based only on a single lot of NanoPOP4. Therefore, I
was unable to assess lot-to-lot variability for anything other than run failure. The results from
the second lot only are discussed here.
Profiler Plus, Identifiler Plus and PowerPlex 16 HS profiles obtained using the
POP-4/36cm array spectral calibrations, coupled with either traditional or experimental run
modules, exhibited very fast separation, but had poor quality profiles, including peak
broadening, poor resolution, spikes, pull-up and additional allele call problems isolated to the
TH01 locus associated with the inability to resolve alleles 9.3 and 10. Subsequent adjustments
including reductions in run voltage and current stability, along with an increase in oven
temperature, did not improve profile quality to an acceptable level. Therefore, spectral
calibrations were performed using NanoPOP4/36cm array prior to any additional testing.
Using default spectral calibration run modules, spectral problems included baseline
separation and various peaks of unknown origin, resulting in failures for numerous capillaries,
more so for dye set F than G5 (for 5-dye life Technologies kits) (see Figure 3.1); dye set D was
78
not tested using default spectral calibration run modules. Even when water was processed
instead of the actual dye sets, peaks were present, indicating a potential interference
originating from the polymer itself. Spectral calibration run modules were altered from the
default to shorten run time, and without any other modifications, the F and G5 dye set
spectrals passed for all capillaries (see Figure 3.2). It is unclear why this small change resulted in
pristine spectral calibrations for all capillaries. It is possible that whatever “contaminant” was in
the system had been flushed out. However, spectral calibrations that were performed
immediately following the successful F and G5 spectrals failed for numerous capillaries using
the same modified run module for Promega matrix standards (arbitrarily name dye set “D”) due
to the presence of extra peaks.
Once spectral calibrations were established for dye sets F and G5 using NanoPOP4/36cm
array, subsequent run module modifications were performed using COfiler and Identifiler Plus
amplification products. The occurrence of pull-up decreased, but other peak quality issues
persisted. Furthermore, a usable allelic ladder was never generated due to the inability to
separate alleles 9.3 and 10 at the TH01 locus, thus nearly all allele calls at that locus were
actually incorrect. Figure 3.3 contains a representative profile obtained using NanoPOP4/36cm
array.
3.1.1.2 NanoPOP6 (22cm and 36cm Array)
Due to the complete inability of electrophoresis from two of the three NanoPOP4 lots
tested, as well as the extremely diminished resolution and inability to obtain usable spectral
calibrations from the third lot, I questioned the quality and applicability of NanoPOP6.
79
A) F Dye Set B) G5 Dye Set C) G5 Dye Set with Water
Figure 3.1. Spectral calibrations obtained using NanoPOP4/36cm array with default POP-4/36cm array modules. The spectral for dye set F (A) exhibits baseline separation and various peaks of unknown origin. The spectral for dye set G5 (B) also exhibits peaks of unknown origin. Even processing water only (no dye set) using the G5 run module (C) resulted in the detection of unexplained peaks, though at lower magnitudes.
A) F Dye Set B) G5 Dye Set
Figure 3.2. Spectral calibrations obtained using NanoPOP4/36cm array with modified run modules. The spectrals for dye sets F (A) and G5 (B) passed for all 16 capillaries.
80
Figure 3.3. Identifiler Plus profile obtained using NanoPOP4/36cm array. Peak broadening and spikes are evident at many loci, as well as non-concordant allele calls at TH01 (and D13) due to poor separation of alleles 9.3 and 10 in the allelic ladder (not shown).
81
Therefore, I decided to post-pone the evaluation of NanoPOP6, at least until after I evaluated
POP-6 on a 22cm array. However, due to the success of the latter polymer/array combination
(see next section), I ultimately abandoned any evaluation of NanoPOP6.
3.1.1.3 POP-6 (22cm Array)
Identifiler Plus STR profiles were processed during the initial evaluation of the
POP-6/22cm array detection method, and three main issues arose with respect to run module
development, including finding the right balance between numerous parameters such that all
necessary data was collected, baseline separation was overcome and good sizing quality of the
internal size standard (e.g., internal lane standard or ILS) was maintained.
It was necessary for data collection to include the start of the primer dye blob (a good
check point to ensure that amplification product was in fact loaded into the well) and all of the
analyzed ILS peaks, which ranged from 75-450bp for the G5 dye set. Default run modules
provided by Life Technologies employ a 60sec pre-run time for 22cm arrays compared to
180sec for 36cm arrays; using a 60sec pre-run time, the dye blob (and quite a bit of
unnecessary data to the left) was captured, but increasing it to either 100sec or 180sec
prevented either the dye blob and/or the first peak (75bp) of the ILS from being collected. Thus,
the 60sec pre-run time was pursued with various lengths of data delay times (1-400sec), from
which a delay of 175sec was chosen in order to begin collection at the correct point with regard
to the dye blob. Furthermore, run time was increased from the traditional POP-4/22cm array
run module time of 720sec to 860sec in order to collect all of the necessary ILS peaks.
82
Various combinations of “voltage number of steps” and “voltage step interval” were
evaluated in order to overcome baseline separation. Using the default POP-4/36cm or
POP-6/36cm array run module combination (40nk and 15sec), baseline separation was a
noticeable problem and continued to be using other combinations (40nk and 20sec, 30nk and
20sec), but baseline issues resolved when the default POP-4/22cm array run module
combination (10nk and 20sec) was evaluated.
Thus, correct timing for data collection and a tighter baseline were achieved using the
above discussed conditions with a 60°C oven temperature (see module #13 in Table A.5), but
examination of the resulting STR profiles revealed a decline in ILS sizing quality compared to
that obtained using traditional POP-4/36cm array detection conditions, several alleles miscalled
as off-ladder (OL) alleles in the larger loci (CSF, D18 and D2), a decrease in precision and a
decrease in resolution (as seen in peak broadening of larger alleles).
ILS sizing quality is a value calculated by the GeneMapper® ID software and is used to
describe the migration quality of a sample based upon the migration of known-sized fragments
of the size standard (Applied Biosystems, 2004). Different oven temperatures were tested (55°C
and 63°C) to compare to 60°C in an attempt to improve ILS sizing quality for POP-6/22cm
detection and demonstrated that higher temperatures produced a sizing quality closer to that
obtained using traditional detection conditions (see Table 3.1 and Figure 3.4). However,
improvements were still needed. It should be noted that use of a 55°C oven temperature
required several additional run module modifications (see module 21 in Table A.5).
The observance of alleles miscalled as OL alleles is another example of a migration-
based problem encountered during the initial evaluation of the POP-6/22cm detection process.
83
Table 3.1
Comparison of Traditional and Alternative Detection Methods
POP-4/36cma
POP-6/22cm
Detection Method 60°C 55°C 60°C 63°C
Total Run Time (min) 39 28 24 24 ILS Sizing Quality
b
Average 95% CI
0.90
[0.89, 0.90]
0.79
[0.79, 0.79]
0.84
[0.84, 0.85]
0.85
[0.83, 0.87]c
Sizing Precision (bp)d
Average Difference 95% CI Observed Maximum
0.036
[0.034, 0.039] 0.18
0.088
[0.083, 0.093] 0.36
0.096
[0.090, 0.10] 0.41
0.092
[0.087, 0.097] 0.35
Peak Height (rfu)g
Average 95% CI
1289
[1263, 1315]
3165
[3102, 3228]
3168
[3105, 3232]
3102
[3038, 3165]
Note. Data obtained from an initial evaluation of alternative detection methods using POP-6 polymer and a 22cm array for detection of Identifiler Plus STR profiles. aData obtained using Cellmark’s established POP-4/36cm detection method, which is based on the traditional HID
POP-4/36cm run module, but has a slightly shorter run time (see Table A.3 in Appendix A). bn=96.
cA larger 95%
confidence interval was observed for this data set due to a poor injection of one sample that resulted in an ILS sizing quality of zero.
dn=615.
gn=3198.
Further investigation revealed that alleles were only miscalled as OLs if an entire plate was
processed (as opposed to one or two injection equivalents of samples) and that the affected
samples had all been subject to “first injection effect” (see Figure 3.5). It is suspected that first
injection effect occurs when the temperature of the oven and/or polymer has not quite
stabilized by the start of the first run and, therefore, causes slight migration delays when
compared to subsequent injections. Thus, the ladder from the first injection of the plate
(positioned in well A01) negatively impacted allele calls for samples in subsequent injections of
the same plate, while later ladders (in wells D06 and H09) negatively impacted allele calls for
samples in the first injection, thereby causing OL allele calls. When samples from the first
injection were only analyzed with the ladder from the same injection (A01), no OLs were
observed. The same was noted when all samples outside of the first injection were only
84
analyzed with ladders after the first injection (D06 and H09). Butler (2012, p. 149-150) discusses
a similar situation related to fluctuations in ambient room temperature.
In response to first injection effect, the three run modules with different oven
temperatures (see modules 13, 14 and 21 in Table A.5) were revisited, each time preheating the
oven to the specified run temperature for a minimum of 15min prior to starting the run.
Previously, the oven was preheated, but was not necessarily held at the run temperature for
15min prior to starting the run. Preheating the oven in this manner eliminated first injection
effect and the associated OL alleles for all three run modules.
Figure 3.4. Internal size standards using traditional POP-4/36cm and alternative POP-6/22cm detection. Three oven temperatures (55°C, 60°C and 63°C) were utilized for POP-6/22cm detection. For all methods, peak broadening increased as peak size (bp) increased, with slightly more broadening observed using POP-6/22cm and higher oven temperatures. For the above samples (top to bottom), ILS sizing quality is 0.90, 0.80, 0.85 and 0.86, respectively.
POP-4, 60°C
POP-6, 55°C
POP-6, 60°C
POP-6, 63°C
85
A) Sample G02 Analyzed With All Three Allelic ladders B) Sample G02 Analyzed With Only Allelic ladder A01
C) Sample B10 Analyzed With All Three Allelic ladders D) Sample B10 Analyzed Without Allelic ladder A01
Figure 3.5. Samples exhibiting first injection effect. Data was obtained from Identifiler Plus, profiles detected using POP-6/22cm array with a 63°C oven temperature (pre-heated to 63°C for <<15min); similar profiles were also obtained using 55°C and 60°C. Allelic ladders were positioned in wells A01, D06 and H09. A sample that was detected during the first injection (sample G02) had OL alleles (circled) at CSF, D2 and D18 when analyzed with all three ladders (A), but when it was analyzed with only the ladder that was processed during the same injection (A01), no OLs were called (B). Similarly, another sample from the same 96-well plate that was detected during a later injection (sample B10) also had OL alleles (circled) at CSF, D2 and D18 when analyzed with all three ladders (C), but when it was analyzed without the ladder from the first injection (A01), no OLs were called (D).
86
Allele sizing precision is crucial for the analysis software to make correct allele calls. Life
Technologies emphasizes that the detection method must be able to correctly size alleles
within ±0.5bp (Applied Biosystems, 2004) and that this can be accomplished using traditional
POP-4/36cm detection, which is advertised as having a standard deviation of <0.15 for allele
sizing precision for alleles up to 250bp (Life Technologies, 2014e). To assess precision, sizing
data was obtained from three Identifiler Plus allelic ladders that were detected using traditional
POP-4/36cm detection and alternative POP-6/22cm detection using three different oven
temperatures (as described above). For all four detection methods, precision tended to decline
as allele size increased (no data shown). Although the average difference and observed
maximum difference in allele sizing roughly doubled for each of the POP-6/22cm detection run
modules as compared to traditional POP-4/36cm detection, no differences greater than
±0.50bp were observed for any method (see Table 3.1), demonstrating acceptable levels of
precision from all based on this criteria.
Resolution was qualitatively assessed by visual inspection of the ILS for peak broadening
(see Figure 3.4) and the ability for the analysis software to differentiate alleles with one base-
pair differences. For all methods, peak broadening increased slightly as peak size (bp) increased,
with slightly more broadening using POP-6/22cm and higher oven temperatures than
traditional POP-4/36cm detection. Unfortunately, the number of loci exhibiting alleles that
differed by one base-pair was extremely small and was limited to three allelic ladders with
alleles 9.3 and 10 at TH01 and one sample with alleles 10.3 and 11 at CSF (see Figure 3.6).
However, the data was still informative and indicated that, similar to peak broadening and
precision, resolution decreased as allele size increased for all four detection methods.
87
A) TH01 B) CSF
Figure 3.6. Resolution using traditional and alternative detection methods. Resolution was qualitatively compared between methods. Alleles 9.3 and 10 (~186-187bp) at TH01 in the allelic ladder (A) exhibit little difference in resolution, but as amplicon length increases, differences in resolution are visible with the naked eye, as seen with alleles 10.3 and 11 (~324-325bp) at CSF from a single source reference sample (B). Resolution decreases with a change from POP-4 to POP-6 and further decreases as temperature increases, such that at 63°C, allele 11 was not differentiated from 10.3 by GeneMapper® ID.
More specifically, resolution differences were (qualitatively) not very noticeable at TH01
between the methods tested, but exhibited larger differences at CSF. In fact, detection with
POP-6/22cm and a 63°C oven temperature resulted in a profile in which alleles 10.3 and 11
could not be differentiated by the analysis software, but were still visually identifiable. The
difference in resolution at these two loci was likely due to the difference in allele size.
Additionally, profiles obtained using POP-6/22cm detection (all three oven
temperatures) exhibited peaks heights that were substantially higher than those obtained using
POP-4, 60°C
POP-6, 55°C
POP-6, 60°C
POP-6, 63°C
POP-4, 60°C
POP-6, 55°C
POP-6, 60°C
POP-6, 63°C
88
traditional POP-4/36cm detection (see Table 3.1), as well as more occurrences of oversaturated
peaks and detectable pull-up peaks (data not shown). Prior to starting this evaluation, it was
suspected that peak height may be notably different between POP-4/36cm and POP-6/22cm
detection, and these data support that POP-6cm/22cm detection yields higher peak heights.
Initial run modules for POP-6/22cm detection utilized 3kV 10sec injections, and given the
increase in peak height, occurrence of oversaturation and pull-up, lower injection voltage
and/or reduced injection time were evaluated further (see section 3.1.2.1 Modifications to
Injection Voltage and Time).
In summary, run module parameters were adjusted during this initial evaluation in order
to collect all necessary data points (from the start of dye blob to all ILS peaks), eliminate
baseline separation and improve migration. Overall, semi-developed POP-6/22cm run modules
utilizing three different oven temperatures yielded data with similar trends as compared to
traditional POP-4/36cm detection: decreased run time; decreased ILS sizing quality; decreased,
yet still acceptable, precision of base-pair sizing of alleles; minimal peak broadening of the
larger amplicons; decreased resolution; increased peak heights; increase in the occurrence of
oversaturation and detectable pull-up (data not shown). Though not perfect, the initial
evaluation of POP-6/22cm array yielded in much more promising profiles than the
NanoPOP4/36cm array and was selected for further optimization.
3.1.2 Optimization of an Alternative Detection Method Using POP-6 (22cm Array)
3.1.2.1 Modifications to Injection Voltage and Time
Using traditional POP-4/36cm detection and each of the five experimental POP-6/22cm
89
run modules at 60°C with varying injection voltages and times, full profiles were obtained from
all Identifiler Plus profiles using a 150rfu threshold. Oversaturation occurred in one sample (1%)
processed with traditional detection, and occurrences of oversaturation in the alternative
POP-6/22cm run modules ranged from 0-27 samples (0-93%), as illustrated in Figure 3.7. As
expected, more samples exhibited oversaturation when higher injection voltages and longer
injection times were used. Additionally, more oversaturation occurred using the same injection
parameters for a POP-6/22cm detection compared to POP-4/36cm detection, which was
expected based upon the increase in peak height noted for POP-6/22cm during the initial
evaluation. Only the 2kV 4sec and 2kV 7sec injections from the POP-6/22cm data sets
generated profiles in which all were free of oversaturation (and maintained full profiles), so
these two were examined further for pull-up and peak height.
Figure 3.7. Detectable pull-up peaks from traditional and alternative detection methods. The percent of alleles generating pull-up peaks using traditional POP-4/36cm detection (with a 3kV 7sec injection, n=94) is compared to that of alternative POP-6/22cm detection using 60°C run modules with various injection voltage and time combinations (n=29, 30 or 31).
0%
20%
40%
60%
80%
100%
2kV 4sec 2kV 7sec 2kV 10sec 2kV 20sec 3kV 7sec
% o
f A
llele
s
Injection Voltage and Time
POP-4/36cm POP-6/22cm
90
The percentage of alleles and samples generating called pull-up peaks ranged from 0.3%
to 3.7% and 10% to 32%, respectively, with the lowest frequencies of pull-up from traditional
POP-4/36cm detection, followed by POP-6/22cm using a 2kV 4sec injection and 2kV 7sec
injection (see Figure 3.8). These findings were statistically significant for POP-4/36cm detection
as compared to POP-6/22cm detection with 2kV 7sec injection (Chi-Square contingency test,
p<0.001 and 0.025<p<0.05 for frequency of alleles and samples with pull-up, respectively). On
the other hand, average percent pull-up (calculated by dividing the peak height of the pull-up
peak by the source peak) was highest with traditional detection (10%), as compared to
POP-6/22cm detection (6% for both injection parameters); these findings were statistically
significant (two-tailed t-test, p=3.02x10-2 and 8.26x10-5 for comparisons of POP-4/36cm
detection to POP-6/22cm with 2kV 4sec and 2kV 7sec injections, respectively). Furthermore,
average peak height of called pull-up ranged from 173rfu to 190rfu for the three data sets,
whereas average allele peak height was 1459rfu, 915rfu and 1440rfu for traditional
POP-4/36cm, POP-6/22cm with 2kV 4sec injection and POP-6/22cm with 2kV 7sec injection,
respectively (see Figure 3.9). Two-tailed match-pair t-tests revealed statistically signficant
differences in allele peak height for both POP-6/22cm data sets compared to traditional
POP-4/36cm detection (p=1.96x10-272 and 3.04x10-2 for 2kV 4sec and 2kV 7sec injections,
respectively). Full profiles were always obtained with peak heights well above the analysis
threshold for POP-6/22cm with 2kV 7sec injection (minimum peak height was 362rfu).
Therefore, despite exhibiting significantly lower peak heights than those obtained from
traditional detection, a 2kV 7sec injection with POP-6/22cm is suitable for the intended use. On
the other hand, the lowest peaks from the 2kV 4sec injection data were close to the analysis
91
threshold (minimum peak height was 220rfu), thereby increasing the risk for potential dropout
when using a 150rfu threshold. Therefore, a 2kV 4sec injection with POP-6/22cm does not yield
as desireable allele peak heights as the 7sec injection.
Figure 3.8. Frequency and percent pull-up from traditional and alternative detection methods. A summary of pull-up peaks obtained using traditional POP-4/36cm detection and alternative POP-6/22cm detection is displayed above, including frequency of alleles generating called pull-up peaks (n=858 or 885), frequency of samples (n=30 or 31) with called pull-up and average percent pull-up.
Figure 3.9. Peak heights from traditional and alterative detection methods. Average peak height of alleles (n=858 or 885) and pull-up peaks (n=3, 4 or 31) are displayed for traditional POP-4/36cm detection and alternative POP-6/22cm detection.
0%
5%
10%
15%
20%
25%
30%
35%
Frequency (Alleles) Frequency (Samples) Average % Pull-up
POP-4/36cm 3kV 7sec POP-6/22cm 2kV 4sec POP-6/22cm 2kV 7sec
0
200
400
600
800
1000
1200
1400
1600
Allele Pull-up
Pea
k H
eigh
t (r
fu)
POP-4/36cm 3kV 7sec POP-6/22cm 2kV 4sec POP-6/22cm 2kV 7sec
92
Of the five alternative POP-6/22cm run modules assessed, the 2kV 7sec module was
selected for further optimization given that full profiles were always obtained without
oversaturation, peak heights were more comparable (than the 2kV 4sec data set) to those
obtained using POP-4/36cm detection and pull-up peak characteristics were similar to (e.g.,
average peak height) and/or offered improvements (e.g., average percent pull-up) compared to
POP-4/36cm, even though the frequency of pull-up was higher than that of traditional
detection.
3.1.2.2 Modifications to Run Voltage, Run Temperature and Current Stability
The POP-6/22cm run modules selected from the previous section for further
optimization exhibited average ILS sizing qualities of 0.84 (60°C) and 0.85 (63°C), which are
lower than that of traditional POP-4/36cm detection (0.89). The eleven run modules elevated
as a result of modifications to run voltage, run temperature and current stability all yielded
profiles with sizing qualities of 0.86-0.90, but with these improvements came longer total run
times (29-36min, compared to 24min from the previous section) required to collect all
necessary data points because the speed of migration had been slowed (see Table 3.2). In
general, a decrease in run voltage from 15kV to 12-14kV had the greatest impact on migration
speed, but also improved ILS sizing quality. Elevation of the run temperature from 60°C to 63°C
increased the speed of migration, as well as increased sizing quality by 0.01, when other
parameters remained the same. Changes in current stability had little to no effect on ILS sizing
quality.
93
Table 3.2
Effects of Run Voltage, Run Temperature and Current Stability on ILS Sizing Quality
Parameters
Run Module
Run Voltage (kV)
Run Temp. (°C)
Current Stability (μA)
Total Run Time (min)
Comments
1 12 60 5 25 Didn’t collect all data
2 12 60 5 36 ILS: 0.89 (all) Can decrease run time
3 13 60 5 36 ILS: 0.88-0.89 Can decrease run time
4 14 60 5 33 ILS: 0.86-0.88, 1 @ 0.57 Can decrease run time
5 14 63 5 32 ILS: 0.87-0.89 Can decrease run time
6 13 63 5 30 ILS: 0.89, 1 @ 0.90 Can decrease run time
7 12 63 5 29 ILS: 0.89, 1 @ 0.90
8 12 60 4 30 ILS: 0.88-0.89
9 12 60 6 30 ILS: 0.88-0.89
10 12 63 4 30 ILS: 0.89-0.9
11 12 63 6 29 ILS: 0.89-0.9
Note. Speed of migration was reduced for all run modules compared to those with a 15kV run current (24min), but ILS sizing quality improved (from 0.84-0.85).
Impacts on precision could not be evaluated based upon the samples that were
processed above, but were evaluated via injections of 12 allelic ladders with POP-6/22cm
detection using seven run modules with oven temperatures of 60°C and 63°C, along with run
voltages of 12-15kV (see Table 3.3). On a side note, average ILS sizing quality from these data
sets ranged from 0.84 (60°C, 15kV) to 0.89 (various 12kV and 13kV modules) and tended to
decrease as oven temperature and voltage increased. As a reminder, traditional POP-4/36cm
detection generated ILS sizing qualities of 0.89-0.90 for the G5 dye set, and since none of the
94
modules were able to achieve this level, I was willing to consider any module that produced ILS
sizing qualities of ≥0.88.
Figure 3.10. Allele sizing differences for traditional and alternative detection methods. Average differences in allele size for traditional POP-4/22cm detection and seven POP-6/22cm run modules are displayed from 12 Identifiler Plus allelic ladders. Differences are grouped by allele size.
Base-pair sizing precision needs to be within 0.5bp in order for GeneMapper® ID to
correctly call alleles, but Life Technologies advertises ±0.5bp precision with a standard deviation
of <0.15 for alleles up to 250bp using traditional POP-4 detection. When data from the 12 allelic
ladder samples was broken down by allele size, average difference in sizing tended to increase
as allele size increased (see Figure 3.10), and two run modules exhibited allele differences in
excess of ±0.50bp (60°C 12kV and 60°C 15kV), but these latter occurrences were limited to a
total of five alleles >300bp. For alleles up to 250bp only, average size difference ranged from
0.037bp to 0.067bp, while average standard deviations ranged from 0.047 to 0.086, with
0.000
0.020
0.040
0.060
0.080
0.100
0.120
0.140
0.160
100-150 150-200 200-250 250-300 >300
Sizi
ng
Dif
fere
nce
(b
p)
Allele Size (bp)
60°C, 12kV 60°C, 13kV 60°C, 15kV 63°C, 12kV
63°C, 13kV 63°C, 14kV 63°C, 15kV POP-4
95
observed maximums of 0.16-0.32bp and 0.076-0.14, respectively (see Table 3.3). Furthermore,
unlike ILS sizing quality, sizing precision was higher for 13kV and 14kV run modules as
compared to 12kV and 15kV and, similar to ILS sizing quality, did improve as oven temperature
increased. Yet, using only the data from alleles up to 250bp, all of the POP-6/22cm run modules
and traditional POP-4/36cm detection would be acceptable based on Life Technologies’
specifications. However, I do not think such a conclusion is sufficiently stringent given that a
substantial portion of STR alleles are greater than 250bp. Furthermore, Promega’s PowerPlex
amplification kits have even larger allele sizes than Life Technologies’ amplification kits.
Therefore, only assessing precision for alleles up to 250bp could be problematic.
Thus, it was more appropriate to examine all allele sizes that were tested and to identify
POP-6/22cm run modules that maintained as good or even better precision than that of
traditional POP-4/36cm detection. As noted earlier, precision decreased as allele size increased
for all methods (see Figure D.1). When all tested alleles were grouped together per run module
(see Table 3.3), average standard deviations ranged from 0.056 to 0.11bp, whereas average size
difference ranged from 0.044bp to 0.085bp. All of these are within Life Technologies’
specifications. However, maximum observed standard deviations and/or allele differences
exceeded their criteria for traditional POP-4/36cm and five of the seven POP-6/22cm run
modules. Compared to traditional POP-4/36cm detection, precision either significantly
improved or was not significantly different using POP-6/22cm 60°C 13kV, 63°C 13kV and 63°C
14kV run modules based on standard deviation (p=8.00x10-7, p=0.468 and p=1.00, respectively,
using the Tukey HSD method) and size differences (p<1.00x10-7, p=3.42x10-4 and p=0.486,
respectively, using the Tukey HSD method) for all allele sizes and not just those up to 250bp.
96
Table 3.3
Effects of Run Voltage and Run Temperature on ILS Sizing Quality and Precision
60°C
63°C
Assessed Value POP-4 12kV 13kV 15kV 12kV 13kV 14kV 15kV
Total Run Time (min) 39 31 27 24 31 27 24 25 ILS Sizing Quality
a
Average 95% CI
0.90
[0.89, 0.90]
0.89b**
[0.89, 0.89]
0.88c***
[0.88, 0.88]
0.84***
[0.84, 0.85]
0.89b
[0.89, 0.90]
0.89b**
[0.89, 0.89]
0.88c***
[0.88, 0.88]
0.87***
[0.87, 0.87] Sizing Precision for All Alleles
d
Standard Deviation Average 95% CI Observed Maximum Size Difference (bp) Average 95% CI Observed Maximum
0.074e [0.070, 0.079]
0.18
0.061h [0.059, 0.063]
0.25
0.10f,g*** [0.095, 0.11]
0.27
0.081k*** [0.078, 0.084]
0.54
0.056*** [0.053, 0.058]
0.13
0.044*** [0.043, 0.045]
0.22
0.11f*** [0.10, 0.11]
0.26
0.085k*** [0.082, 0.088]
0.51
0.089g*** [0.084, 0.093]
0.21
0.073j*** [0.071, 0.075]
0.33
0.068e* [0.064, 0.071]
0.16
0.055i*** [0.053, 0.057]
0.29
0.073e [0.070, 0.076]
0.14
0.058h,i* [0.056, 0.060]
0.25
0.094g*** [0.090, 0.099]
0.226
0.074j*** [0.072, 0.074]
0.44 Sizing Precision up to 250bp
l
Standard Deviation Average 95% CI Observed Maximum Size Difference (bp) Average 95% CI Observed Maximum
0.055m [0.050, 0.061]
0.094
0.045r [0.043, 0.047]
0.16
0.075o*** [0.067, 0.084]
0.13
0.059s*** [0.055, 0.063]
0.30
0.047*** [0.043, 0.051]
0.076
0.037*** [0.035, 0.039]
0.16
0.086q*** [0.078, 0.094]
0.14
0.067*** [0.063, 0.071]
0.32
0.074o,o*** [0.067, 0.080]
0.13
0.060s*** [0.057, 0.063]
0.20
0.058m,n [0.053, 0.062]
0.096
0.047r [0.045, 0.049]
0.18
0.066n,o*** [0.061, 0.070]
0.13
0.041*** [0.039, 0.043]
0.24
0.078o,q*** [0.072, 0.085]
0.12
0.061s*** [0.058, 0.064]
0.27
Note. Data obtained using traditional POP-4/36cm and alternative POP-6/22cm detection for 12 Identifiler Plus allelic ladders. Averages sharing a common subscript are not statistically different at α=0.05 according to the Tukey HSD procedure. Two-tailed t-tests were performed for comparison of each POP-6/22cm run module to traditional POP-4/36cm. an=12.
dn=2460.
ln=1428.
*p < 0.05. **p < 0.01. ***p < 0.001.
97
Of these three POP-6/22cm run modules, observed maximum standard deviation and allele
differences did not exceed 0.15 and 0.50bp, respectively, for the 60°C 13kV and 63°C 14kV run
modules only.
While taking ILS sizing quality, precision and total run time from the above seven run
modules into consideration, two modules were selected for further evaluation: 60°C 13kV and
63°C 14kV. The 60°C 13kV module had a total run time of 27min and, as stated previously,
offered slightly reduced ILS sizing quality, but significantly improved precision compared to
POP-4/36cm. In contrast, the 63°C 14kV module had a 24min run time and also had slightly
reduced ILS sizing quality (though not significantly different from the 60°C 13kV module) with
comparable precision to POP-4/22cm detection. Of these two POP-6/22cm run modules, I
decided that the improved precision exhibited by the 60°C 13kV module did not outweigh its
longer detection time, especially given that the 63°C 14kV run module yielded as good precision
as that of POP-4/36cm. Therefore, several detection plates were processed using the 63°C 14kV
run module, and those results were used for its validation (see validation section below).
3.1.3 Validation of an Alternative Detection Method Using POP-6 (22cm Array)
In this and subsequent sections, “POP-4” refers to traditional, HID POP-4 detection on a
36cm array and “POP-6” refers to the alternative POP-6 detection method developed on a
22cm array. See Figures D.3 – D.6 for representative electropherograms.
3.1.3.1 Comparison to Traditional POP-4/36cm Detection
Upon review of the validation data, two capillaries (13 and 14) were identified from the
98
22cm array that were not functioning properly; this can occur when an array is used
infrequently and is not considered to be indicative of a defect in the POP-6/22cm platform. All
data from these two capillaries were removed from the POP-6 data sets prior to analysis.
3.1.3.1.1 Reproducibility (Concordance of Allele Calls)
For Identifiler Plus samples, all alleles detected using both methods were concordant
with one another, except for one out-of-bin “OL” allele at FGA detected with POP-4 that was
called a 26 with POP-6. Subsequent reinjection of this sample using POP-4 resulted in the OL
allele being called a 26. There was an additional OL at D19S433 detected with POP-4 and POP-6
that was given an allele call of “<9” during GeneMapper® ID analysis for both detection
methods.
For PowerPlex 16 samples, all alleles detected using both methods were concordant
with one another. There was one OL at D16S539 detected with POP-4 that could not be
confirmed from its original data set because the corresponding sample processed with POP-6
exhibited migration problems (rendering that particular profile unusable), but it was confirmed
as a micro-variant OL (13.3) from a previous POP-4 injection.
3.1.3.1.2 Resolution
A limited number of samples (n=1 for Identifiler Plus at TH01 and n=2 for PowerPlex 16
at TH01/D7) were identified from the total of 356 that were processed during this validation
that exhibited alleles with one base-pair differences. In addition to these samples, all six allelic
ladders from each of the amplification types (using alleles 9.3 and 10 at the TH01 locus)
99
underwent a resolution assessment using the valley value calculation (see Table 3.4). The
average valley values were 0.447 and 0.448 from Identifiler Plus and 0.359 and 0.368 from
PowerPlex 16 using POP-4 and POP-6 detection, respectively (not significantly different using a
two-tailed, match paired t-test, p=0.949 and 0.538, respectively). It should be noted that an
increase in valley value corresponds with a decrease in resolution (see Figure D.2).
Furthermore, this assessment may lead to an incorrect conclusion that the degree of
one base-pair resolution is different from Identifiler Plus alleles as compared to PowerPlex 16,
when in fact the observed difference may be more likely attributable to the size of the alleles
assessed in each case. All of the assessed Identifiler Plus alleles were from TH01 alleles 9.3 and
10 and had base-pair sizes of 186.13-187.23bp (based on POP-4 data); on the other hand,
PowerPlex 16 data originated from TH01 alleles 9.3 and 10 (n=7) and D7 (n=1), with base-pair
sizes of 175.21-176.27bp and 227.43-228.45bp, respectively. With this is mind, it should also be
noted that resolution of the D7 alleles (Vv=0.559 using POP-6) was poorer than that of the TH01
alleles (average Vv=0.343 using POP-6), which could be explained by the larger size of the D7
alleles.
To test this theory, resolving power, or resolution length (RSL), was calculated for alleles
in the allelic ladder. Linear regression analysis demonstrated that resolution decreased (i.e., the
RSL value increased) for both allelic ladders using both detection methods as allele size
increased (see Figure 3.11). Furthermore, the rates of decreasing resolution (as indicated by the
slopes of the regression lines) were significantly different between POP-4 and POP-6 detection
(p=0.0338 and p=2.39x10-6, two-tailed t-test for Identifiler Plus and PowerPlex 16, respectively),
as were their elevations (as indicated by the y-intercepts of the regression lines; p<<0.0001,
100
two-tailed t-test for Identifiler Plus and PowerPlex 16). It should also be noted that the linear
regression lines illustrate that POP-6 has higher resolution than POP-4 for smaller alleles
(<200bp), but that the opposite is true for larger alleles. In fact, POP-6 resolution length
exceeds the desired threshold (RSL 1) for alleles larger than ~300bp, which could be
problematic if alleles of this size (or greater) differ by one base-pair at the same locus.
A) Identifiler Plus
B) PowerPlex 16
Figure 3.11. Resolving power of traditional and alternative detection methods. Linear regression analysis from an Identifiler Plus allelic ladder (n=186 alleles) is displayed on top (A), while that from a PowerPlex 16 allelic ladder (n=192 alleles) is on bottom (B). For POP-4 and POP-6 detection, as allele size (bp) increases, resolution length increases, indicating a decrease in resolution. POP-6 offers higher resolution than POP-4 for smaller alleles (<200bp), but lower resolution for larger alleles (>200bp).
POP-4 y = 0.0010x + 0.4859
R² = 0.7019
POP-6 y = 0.0018x + 0.3111
R² = 0.6928 0.0
0.2
0.4
0.6
0.8
1.0
1.2
0 50 100 150 200 250 300 350 400
RSL
Allele Size (bp)
POP-4 POP-6 Linear (POP-4) Linear (POP-6)
POP-4 y = 0.0006x + 0.6161
R² = 0.345
POP-6 y = 0.0019x + 0.3522
R² = 0.6161
0.0
0.5
1.0
1.5
2.0
0 50 100 150 200 250 300 350 400 450 500
RSL
Allele Size (bp)
POP-4 POP-6 Linear (POP-4) Linear (POP-6)
101
A) Identifiler Plus
B) PowerPlex 16
Figure 3.12. ILS migration for traditional and alternative detection methods. Representative ILS migration for Identifiler Plus and PowerPlex 16 is displayed, with traditional POP-4 detection on top and POP-6 on bottom for A and B. (A) LIZ is the internal size standard used with the Life Technologies 5-dye amplification kits, including Identifiler Plus. The sizing quality of the POP-4 sample shown is 0.90, while that of the POP-6 sample is 0.88, which are typical sizing qualities seen for these two polymers. (B) ILS600 is the internal size standard used with the Promega 4-dye amplification kits, including PowerPlex 16. The sizing quality of the POP-4 sample shown is 1.0, while that of the POP-6 sample is 0.99, which are typical sizing qualities seen for these two polymers.
3.1.3.1.3 ILS Sizing Quality
ILS sizing quality was collected for 192 injections using POP-4 and 168 injections using
102
POP-6 for Identifiler Plus and PowerPlex 16 profiles. On average, passing sizing qualities were
slightly higher for POP-4 (0.90 for Identifiler Plus and 1.0 for PowerPlex 16) than POP-6 (0.88 for
Identifiler Plus and 0.97 for PowerPlex 16) for both kits, and these differences were statistically
significant (p=1.4x10-135 for Identifiler Plus and p=1.3x10-27 for PowerPlex 16; see Table 3.4).
Despite being significant, these differences did not result in adverse effects, such as excessive
OL alleles due to poor migration, miscalled alleles or inability to resolve alleles. Representative
electropherograms of ILS migration are in Figure 3.12.
3.1.3.1.4 Peak Height
Average peak height ranged from 990rfu to 1103rfu and was significantly lower using
POP-6 for both types of amplification product (two-tailed t-test, p=6.81x10-23 for Identfiler Plus
and p=8.65x10-18 for PowerPlex 16; see Table 3.4). Despite this significant difference, average
peak height was within the desired range (750-1500rfu) for all data sets, but observed
minimums were right at the analysis thresholds, which is not favorable because it increases the
potential for allelic dropout. Sensitivity range is discussed in more detail below.
3.1.3.1.5 Pass Rates
Identifiler Plus and PowerPlex 16 pass rates were lower for POP-6 (87% and 84%,
respectively) than POP-4 (93% and 88%, respectively) generally due to allele peak heights being
slightly below threshold (for Identifiler Plus) or injection failures (see Table 3.4). However, these
differences were not significant (two-tailed t-test). The most substantial difference appears to
be injection failures, but as discussed previously, the instrument setup with POP-6/22cm array
103
was used infrequently, which can increase injection failures. Thus, this difference may be
explained as an instrument usage issue and not related to the protocols themselves.
Table 3.4
Comparison Between POP-4 and POP-6 Detection
Identifler Plus
PowerPlex 16
Assessed Value POP-4 POP-6 POP-4 POP-6
Total Run Time (min) 39 24 45 28 Valley Value
a
Average 95% CI
0.447
[0.423, 0.472]
0.448
[0.417, 0.480]
0.359
[0.326, 0.391]
0.368
[0.301, 0.435] ILS Sizing Quality
b
Average 95% CI
0.90*
[0.90, 0.90]
0.88*
[0.88, 0.88]
1.0**
[1.0, 1.0]
0.97**
[0.97, 0.98] Peak Height (rfu)
c
Average 95% CI Observed Maximum Observed Minimum
1103***
[1087, 1118] 4810 181
991***
[975, 1007] 4444 180
1080****
[1066, 1095] 4280 150
990****
[976, 1005] 4327 150
Pass Ratesb
Dropout PHR <50% Pull-up (>20%) Stutter (>20%) OL Injection Failure
93% 1.1% 4.4% 0% 0%
1.1% 0.56%
87% 5.1% 4.5% 0% 0%
1.3% 2.6%
88% 4.4% 3.9%
0.56% 0.56% 1.7% 1.1%
84% 5.1% 5.1% 0%
0.63% 1.3% 3.8%
Sizing Precisiond
Standard Deviation Average 95% CI Observed Maximum Sizing Difference (bp) Average 95% CI Observed Maximum
0.065 [0.062, 0.068]
0.13
0.052 [0.050, 0.054]
0.25
0.047 [0.045, 0.049]
0.12
0.037 [0.036, 0.038]
0.20
Note. Data obtained using traditional POP-4/36cm and alternative POP-6/22cm detection. an=7 for Identifiler Plus and n=8 PowerPlex 16.
bn=180, 156, 180 and 158, respectively.
cn=5128, 4449, 5068 and
5031, respectively. dn=2255 for Identifiler Plus and 2310 for PowerPlex 16 HS.
*p=1.4x10-135
(two-tailed t-test). **p=1.3x10-27
(two-tailed t-test). ***p=6.81x10-23
(two-tailed t-test). ****p=8.65x10
-18 (two-tailed t-test).
3.1.3.2 Precision
Using 11 allelic ladders from Identifiler Plus and PowerPlex 16 HS, average standard
deviation for allele sizing was 0.065 and 0.047, respectively (see Table 3.4). The observed
104
maximum standard deviations were <0.15 for both kits. Furthermore, average allele sizing
differences were 0.052bp and 0.037bp, respectively, and never exceeded 0.50bp. Thus,
precision met the requirements with regard to standard deviation (<0.15) and sizing difference
(<0.5bp) for both amplification kits using POP-6 detection that Life Technologies established for
traditional POP-4 detection.
Furthermore, it should be noted that allele sizing differences increased as allele size
increased for both kits (see Figure 3.13). For Identifiler Plus alleles, the following allele sizes did
not exhibit statistically significant differences in sizing difference (using the Tukey HSD method,
α=0.05): 100-150bp and 150-200bp; and 200-250bp and 250-300bp. Both of these ranges (100-
200bp and 200-300bp) had significant differences between each other and the >300bp range
(p<4.79x10-3 for all). For PowerPlex 16 HS alleles, the following allele sizes did not exhibit
statistically significant differences in sizing difference (using the Tukey HSD method, α=0.05):
100-150bp, 150-200bp, 200-250bp and 250-300bp; 300-350bp and 350-400bp; and 350-400bp
and >400bp. These ranges had significant differences between each other (p<2.55x10-2 for all).
Figure 3.13. Allele sizing differences for Identifiler Plus and PowerPlex 16 HS. Average allele sizing differences (bp) are displayed by allele size for each kit using the validated POP-6 detection method. Eleven allelic ladders were processed from each kit, with samples sizes ranging from 275 to 517 for each allele size. aFor Identifiler Plus, includes all alleles >300bp (up to ~360bp).
0.000
0.020
0.040
0.060
0.080
100-150 150-200 200-250 250-300 300-350 350-400 >400
Sizi
ng
Dif
fere
nce
(b
p)
Alelle Size (bp)
Identifiler Plus PowerPlex 16 HS
a
105
Table 3.5
Summary of Validated POP-6/22cm Injection Times
Identifler Plus
PowerPlex 16
Assessed Value 2kV 4sec 2kV 7sec 2kV 15sec 2kV 4sec 2kV 7sec 2kV 15sec
Peak Height (rfu)a
Average 95% CI Observed Maximum Observed Minimum
416
[387, 445] 1975 181
843
[789, 897] 4045 180
1235
[1147, 1323] 6481 190
365
[340, 390] 1376 153
619
[576, 661] 2504 158
1146
[1067, 1226] 4628 209
Pass Ratesb
Dropout 14% 86%
93% 7.1%
100% 0%
0%
100% 69% 31%
100% 0%
Pass Rates (75rfu)b
Dropout 86% 14%
100% 0%
100% 0%
85% 15%
100% 0%
100% 0%
Note. Data obtained using alternative POP-6/22cm detection with analysis thresholds of 180rfu (Identifiler Plus) and 150rfu (PowerPlex 16), unless otherwise noted. an=192-338.
bn=14 for Identifiler Plus and 13 for PowerPlex 16.
3.1.3.3 Sensitivity and Usable Injection Times
For both Identifiler Plus (n=14) and PowerPlex 16 (n=13) amplification product, one
injection worth of samples from the POP-4/POP-6 comparison study was injected again using
POP-6 with 2kV 4sec and 2kV 15sec injections. As expected, average peak height decreased as
injection time decreased, ranging from 416-1235rfu and 365-1146rfu for Identifiler Plus and
PowerPlex 16, respectively (see Table 3.5). Additionally, more allele dropout occurred as
injection time decreased, such that 86% and 100% of samples experienced dropout using a 2kV
4sec injection for Identifiler Plus and PowerPlex 16, respectively, but no dropout was observed
with 2kV 15sec injections. No oversaturation, pull-up (>20%) or elevated baseline were
observed for any profiles. A single, passing pull-up peak ( 20%) was observed using a 2kV 4sec
injection of PowerPlex 16 and a 2kV 15sec injection of Identifiler Plus profiles. Decreasing the
analysis threshold to 75rfu resulted in the detection of full profiles from 86% and 85% for
Identifiler Plus and PowerPlex 16, respectively, using a 2kV 4sec injection and 100% full profiles
106
from all other samples. However, occurrences of pull-up and stutter (all non-failing) also
increased (no data shown), but elevated baseline was not present. Thus, 2kV 4-15sec injections
are suitable for POP-6 detection, but dropout is more likely to occur using 4-7sec injections
when 150rfu analysis thresholds are utilized. Therefore, I suggest using thresholds of 75rfu to
obtain substantially more full profiles, despite a slight increase of detectable – though non-
failing – pull-up and stutter peaks.
3.2 Development of Fast PCR Protocols for Identifiler, Identifiler Plus and PowerPlex 16 HS
3.2.1 Initial Evaluation
3.2.1.1 AmpliTaq Gold® Fast PCR Master Mix
The Identifiler amplifications processed with AmpliTaq Gold® Fast PCR Master Mix
exhibited allelic dropout, less than desired peak heights, inter-locus peak imbalance (i.e,
perferiential amplification) and -A (most notably at TH01 and vWA)(see Figure 3.14). No signs of
amplification were seen in the negative control. As a reminder, this product was not developed
by the manufacturer for multiplexing, but cycling lengths were extended for denaturation,
annealing and extension to help compensate (see Table 2.4). It theorized that allelic dropout,
low peak height and locus imbalance could be a combination of several factors, including the
need for additional dNTPs, increased annealing and/or extension times, modifications to primer
annealing temperature, increased cycle number, modification to template DNA amount, etc. It
was suspected that incomplete adenylation (-A) could be resolved by increasing final extension
time and/or potentially changing the final hold to 25°C instead of 4°C (Foster & Laurin, 2012).
107
3.2.1.2 KAPA2G™ Fast Multiplex PCR Kit
The Identifiler amplifications processed with the KAPA2G™ Fast Multiplex Kit resulted in
full profiles with good peak heights and balance, but -A and some elevated stutter (>10%) were
noted (see Figure 3.14), but overall, yielded better profiles than the other products. No signs of
amplification were seen in the negative control. As a reminder, this product was the only one
evaluated that was developed by the manufacturer for fast PCR and multiplexing, and it also
had the longest PCR thermal cycler program (~50min compared to ~30min for the other
products). As mentioned above, -A could likely be eliminated by increasing the final extension
time. I suspected that elevated stutter could be addressed by altering template amount.
3.2.1.3 SpeedSTAR™ HS DNA Polyermase
The Identifiler amplifications processed with the SpeedSTAR™ HS DNA Polyermase
exhibited decent peak height and balance, but also showed non-specific amplification, elevated
stutter (>10%) and elevated background (see Figure 3.14). No signs of amplification were seen
in the negative control. I anticipated that modifications to the primer annealing temperature of
±1-2 C would improve specificity without adversely affecting the rest of the profile. Changes to
template amount could aid in elevated stutter and background. Additionally, this polymerase
was supplied with a second buffer, Fast Buffer II, which could have been elevated along with
Fast Buffer I.
3.2.1.4 Type-it Microsatellite PCR Kit
None of the Identifiler amplifications processed with the Type-it Microsatellite PCR Kit
108
resulted in the detection of alleles above 75rfu, but some peaks below threshold were
observed (see Figure 3.14). As a reminder, this product was not developed by the manufacturer
for fast PCR, even though it utilizes an enzyme capable of fast PCR, the HotStarTaq® Plus DNA
Polymerase. Additionally, the thermal cycling parameters given for an alternate product (the
Fast PCR Cycling Kit) from the same manufacturer utilizing the same polymerase were used in
an effort to compensate for this product’s lack of fast PCR optimization, but the reduced cycling
times did not work as hoped. I suspected that numerous modifications (primarily, increasing
cycling time) would need to be employed in order for this kit to successfully perform fast PCR.
3.2.2 Development of Four Fast PCR Protocols for Identifiler
Unless otherwise noted, all allele calls were concordant with the expected profile and all
controls performed as expected.
3.2.2.1 Identifiler Fast PCR with the KAPA2G™ Fast Multiplex PCR Kit
As identified during the initial evaluation, the main obstacle to overcome with fast PCR
using KAPA2G was the presence of -A. This was resolved by increasing the final extension time
without increasing PCR time any more than necessary. As expected, the presence of -A
decreased as template amount decreased and as final extension time increased. For the 0.25-
1.0ng DNA series amplified with a 2min final extension, -A was detected using 0.50ng, and
small -A shoulders were still present at the 0.25ng level, at which point, allelic dropout began to
occur (data not shown). For the series amplified with a 5min final extension, -A was still present
using 1.0ng, though to a lesser degree than those amplified with the 2min final extension.
109
A) AmpliTaq Gold Fast B) KAPA2G
C) SpeedSTAR D) Type-it
Figure 3.14. Representative profiles from initial evaluation of four fast PCR methods. All calls made by GeneMapper® ID at a 75rfu threshold are displayed.
110
However, amplification of 0.25-0.75ng yielded STR profiles with no signs of -A while still
maintaining full profiles, but peak heights decreased and locus imbalance increased using
0.25ng. Since -A was increasing with template amount, I was interested in observing how -A
would behave as the fast PCR development process continued without yet increasing final
extension any further.
Thermal cycling parameters were then modified to reduce the overall amplification time
as much as possible without adversely affecting profiles completeness, peak heights, inter- and
intra-locus balance, stutter, pull-up, etc. Starting with a 2-step PCR cycle (with 30sec combined
annealing/extension) and assessment of three primer annealing temperatures (59°C, 61°C and
63°C), a general decrease in peak height was noted compared to 3-step PCR cycling. Full profiles
were obtained from all 63°C samples (0.50-1.25ng); a single allele dropped out using 0.50ng
amplified with 61°C, and allelic dropout occurred for all but 0.75ng with 59°C. As expected, TPH
and average peak heights tended to increase as DNA template increased, and in general, peak
heights were higher for amplifications performed with a 61°C annealing/extension temperature
than those at 59°C or 63°C. Looking at inter-locus imbalance as a whole (not necessarily as a
result of preferential amplification of smaller loci) using the coefficient of variation of each
profile’s locus peak height to total sum of peak heights (LPH:TPH) ratio, inter-locus peak
imbalance was more prominent for the 59°C data set compared to 61°C and 63°C, as indicated
by higher LPH:TPH CV values (see Figure 3.15). Overall, there was a general increase in
imbalance as template amount increased, which was associated with preferential amplification
of smaller loci. I suspected this would improve with a longer annealing/extension. Instances of
PHR <50% were greatest from the 59°C data set (four occurrences in three profiles), compared
111
to single instances of PHR <50% from the other two annealing/extension temperatures. All
stutter peaks remained under 20%, -A was not present (despite being present using 1.00ng DNA
during the previous study) and pull-up and elevated background noise were minimal, if at all
present. No signs of NSA were observed.
The 59°C data set clearly resulted in the poorest quality profiles, which was somewhat
surprising given that the standard Identifiler amplification utilizes a 59°C annealing
temperature, but the protocol under development utilized different amplification chemistry
and employed a 2-step, not 3-step, PCR cycle. The 61°C and 63°C amplifications resulted in
similar quality profiles. Despite a single dropout event with 61°C (and none with 63°C), the 61°C
combined annealing/extension temperature was selected for further testing because that data
set exhibited the highest peak heights.
Figure 3.15. Effects of annealing temperature on inter-locus balance for Identifiler fast PCR with KAPA2G. The coefficient of variation for a profile’s LPH:TPH is higher for amplifications performed at 59°C, but is about the same for those performed at 61°C and 63°C, indicating greater inter-locus peak imbalance at 59°C (n=1 for each data set). Amplification of 1.00ng at 59°C was excluded due to extensive dropout.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
1.25ng 1.00ng 0.75ng 0.50ng 9947A
CV
of
LPH
:TP
H
59° C 61° C 63° C
112
Continuing on with a 5min final extension and 61°C combined annealing/extension step
(30sec), I next examined a decrease in denaturation time (from 15sec down to 10sec and 5sec)
prior to moving forward with an increase in the length of the annealing/extension step (to
improve inter-locus balance). Full profiles from all samples were only obtained from the 5sec
data set, while the other two with longer denaturation times exhibited a single dropout event
each (both using 0.5ng DNA). Peak height, inter-locus and intra-locus balance did not
significantly differ between denaturation times. All three data sets exhibited one PHR <50%. No
called -A was present, but -A peaks below threshold were observed; no concerning stutter, pull-
up, NSA or baseline issues were observed. Thus, to maximize reduction in amplification time,
the 5sec denaturation was selected.
Next, the combined annealing/extension step was increased from 30sec to 35sec and
40sec in an attempt to improve inter-locus balance and increase peak heights to prevent allelic
dropout. Full profiles were obtained from all data sets, but peak heights nearly doubled using a
40sec annealing/extension step (averaged 1220rfu) compared to 30sec and 35sec (averaged
695rfu and 676rfu, respectively). Average PHR improved to 86.7% with 40sec
annealing/extension, which was an increase from 82.9% (30sec) and 82.6% (35sec) for the
other two data sets, but was not significant (p=0.0715, one-way ANOVA). PHR <50% were not
observed for the 35sec and 40sec data sets, but a single PHR was <50% for the 30sec data set.
As predicted, general inter-locus balance tended to improve slightly as annealing/extension
increased, but was not significant (p=0.114, one-way ANOVA). Three instances of detected -A
were present, one in each of the data sets and were limited to the TH01 and vWA loci; other -A
peaks at these loci were also noted below threshold. No other noteworthy findings regarding
113
additional artifacts were observed. From this study, the 40sec annealing/extension time was
selected for additional testing.
Additional efforts to reduce total amplification time included a reduction in initial
activation time by evaluating 1min and 2min activations in comparison to 3min. Full profiles
were obtained from all samples, with similar peak heights, inter- and intra-locus balance. No
called -A was observed, but minimal -A peak morphology below threshold was present. Thus,
the 1min initial activation time was selected for further testing.
Initial efforts to eliminate -A settled on a 5min final extension, but -A was occasionally
observed either above or below threshold as additional testing continued. Further efforts to
address -A were postponed to prevent unnecessary elongation of the total amplification time,
but once all other issues were resolved and amplification time had been reduced as much as
possible, -A was once again revisited. Final extension times of 3min and 4min were elevated in
comparison to the 5min that was used during most of the fast PCR development work to
determine if a shorter final extension could be employed without causing detectable -A.
However, as expected, detectable -A peaks were present using both of the shorter extension
times, more so with the 3min final extension. Using a larger data set from different individuals
(n=24), 5min final extension also proved to be insufficient. Thus, a 10min final extension was
examined with this larger data set and was successful in eliminating all -A (detected and
otherwise).
Completion of the fast PCR development using KAPA2G with the Identifiler primer set
resulted in a 43min amplification protocol, which is about one quarter of the time needed to
perform standard Identifler amplification on the same thermal cycler (see Table 2.6). Based on
114
the data that utilized the selected protocol (n=24), full profiles were obtained from all samples
with an average peak height of 821rfu, an average PHR of 86.8%, 4% of samples exhibiting a
single PHR <50%, an average LPH:TPH CV value of 0.294 and no -A or NSA. Comparison of
profile quality obtained from this protocol to those from standard Identifiler amplification and
the other three fast PCR protocols is discussed in the comparison section below.
3.2.2.2 Identifiler Fast PCR with the AmpliTaq Gold® Fast PCR Master Mix
Based on the initial evaluation of AmpliTaq Gold® Fast PCR Master Mix, profiles
exhibited allelic dropout, less than desired peak heights, inter-locus peak imbalance, -A (most
notably at TH01 and vWA), but no non-specific amplification. As with fast PCR development
with KAPA2G, development began with eliminating -A through longer final extension times.
Initially, a 1min final extension was evaluated (recommended by the manufacture) and this was
compared to 5min and 10min final extensions. Similar to KAPA2G, -A was most prominent using
a 1min final extension, was not called using 5min (though -A peak morphology was observed
below threshold in some samples) and was not observed with a 10min final extension. Based on
these observations and findings from KAPA2G development, I selected the 10min final
extension.
Next, increasing peak height to prevent allelic dropout was pursued in conjunction with
2-step PCR cycling utilizing a 50sec combined annealing/extension step at three different
temperatures (59°C, 61°C and 63°C). Though no signs of NSA had been observed for this
product, I wanted to test different temperatures since a 2-step PCR cycle was under evaluation.
More full profiles were obtained from the 61°C data set (78%) than 59°C (56%) and 63°C (0%).
115
Profiles obtained using 63°C exhibited dropout at D7, D13 and FGA, which are not all large loci,
so I concluded that dropout for this data set was more related to decreased primer annealing
ability and less so from too short of an annealing/extension time (which I attributed as a
possible cause for dropout with the other two data sets). Interestingly, full profiles were only
consistently obtained from 0.50ng amplifications from the 59°C data set, and 0.25ng and
0.75ng inputs exhibited dropout. Therefore, a 61°C combined annealing/extension temperature
was the clear choice to further evaluate from this study. It should be noted that low PHR (<50%)
were observed in two-thirds of the samples from the 61°C this data set, as well as a significant
portion of samples from the other data sets.
Therefore, improvements in PHR were attempted via various denaturation (5sec and
10sec) and combined annealing/extension (40sec, 50sec and 60sec) times. From the six data
sets tested, only the 10sec/60sec yielded full profiles from all samples. Additionally, average
peak heights ratios were acceptable for all amplification parameters tested, with a slight
increase in PHR as template amount increased. However, the occurrence of PHR <50%
continued to a problem and was not completely eliminated for any data set. The number of
PHR <50% decreased as input DNA increased and denaturation and/or annealing/extension
times increased. A closer examination of the loci in which PHR were <50% revealed that one
particular locus (D18) was consistently a problem for all replicates of one of the two samples
tested. Not only is D18 one of the largest loci obtained from the Identifiler primer set, but the
particular alleles at this locus (16 and 22) are about 24bp different in length (size), which was
the largest difference between two alleles at any given loci for the samples tested.
116
This prompted me to examine PHR closer with regard to locus size and allele size
differences; two important trends were noted for heterozygous loci. First, the larger sized allele
(bp) tended to have a lower peak height than the smaller allele, and when PHR were <50%, 92%
of the time it was the larger sized allele that had a lower peak height, compared to 68% of the
time when PHR >50%. Second, alleles exhibiting size differences of 20bp (or allele repeat
differences of 5) had acceptable PHR (>50%) ≥96% of the time. However, one sample was
identified with a 24bp difference (allele repeat difference of six) between its alleles at D18 and
only exhibited acceptable PHR at this locus 7% of the time. Therefore, I wanted to make sure
that samples exhibiting large allele differences were tested for all fast PCR protocols.
Efforts to improve PHR continued via examination of longer denaturation times of 15sec
and 20sec, as suggested by previous studies (Giese et al., 2009; Walsh, Erlich, & Hilguchi, 1992),
but were unsuccessful. Therefore, use of the standard Identifiler thermal cycling protocol was
evaluated to ensure that acceptable PHR could be obtained using this polymerase. As a result,
PHR did improve signficantly for the locus/sample in question, with 89% of the replicates
yielding PHR >50%. Despite this improvement, the sample itself may have been prone to low
PHR at that locus. Nonetheless, efforts continued to improve PHR via a 2min
annealing/extension step (the sum of the separate annealing and extension steps used for
standard amplification), and though PHR for D18 did improve compared to the 1min
annealing/extension previously tested, two-thirds of the replicates had PHR <50% for each of
the DNA inputs (0.25-0.75ng). When denaturation was increased to 1min, along with 2min
annealing/extension, two-thirds of the samples again exhibited PHR <50%. Furthermore, using
117
such long denaturation and annealing/extension steps also increased occurrences of stutter,
pull-up and -A.
At that point, I decided to diverge from 2-step PCR cycling back to 3-step, with a thermal
cycling protocol similar to that used for standard Identifiler amplification, by employing
separate 1min denaturation, annealing (61°C) and extension (68°C and 72°C) steps. Using a
68°C extension, full profiles were obtained from all samples, but two (22%) had PHR <50%.
Using a 72°C extension, all PHR were >50%, but one (11%) of the 0.25ng replicates experienced
a single allele dropout. No additional noteworthy differences were noted between the two
extension temperatures, including average peak height and average PHR, which indicated that
dropout and PHR <50% are just as likely from one extension temperature compared to the
other. Therefore, the 68°C was chosen for further testing, but 72°C would likely have sufficed as
well.
Next, efforts focused on reducing the 3-step PCR cycling protocol such that PHR
remained >50%, beginning with a reduction in denaturation time from 1min back down to
10sec, from which full profiles were obtained from all samples and all PHR were >50%. Chosing
to move forward with a 10sec denaturation, I evaluated reduced annealing times (1min
compared to 45sec, 30sec and 15sec) next; as expected, peak heights decreased and allelic
dropout increased as annealing time and DNA input decreased, with allelic dropout occurring
for all DNA amounts using a 15sec annealing. Though full profiles were obtained for 30sec
annealing, average peak height was low (502rfu), making it more succeptable to dropout than
the 45sec and 1min annealing times, which had average peak heights of 946rfu and 1248rfu,
respectively. Average PHR did not differ between the annealing times, but occurrences of PHR
118
<50% were observed for 45sec and 30sec using 0.25ng DNA only, but were present using as
much as 0.75ng DNA with a 15sec annealing time. Inter-locus peak balance also decreased as
annealing time decreased; balance was acceptable for all but the 15sec data set. Thus, the 1min
and 45sec annealing times were most suitable for selection, and 45sec was chosen because it
offered a reduction in amplification time. The last attempt at reducing total amplification time
was through shorter extension times (1min compared to 45sec and 30sec). Full profiles were
obtained from all samples, but peak heights decreased for 30sec extension compared to 45sec
and 1min, as well as for decreases in DNA input, such that the 30sec extension time was ruled
out (average peak height of 636rfu). The 1min and 45sec extension data sets exhibited similar
average peak heights (946rfu and 1024rfu, respectively). Inter- and intra-locus were similar for
all three data sets, and PHR <50% were always limited to amplification of 0.25ng DNA. Thus, a
45sec extension step was selected.
It should also be noted that from these data sets, -A was occasionally observed, likely
due to increases in amplification time compared to fast PCR protocols utilized early on in the
development process. Two additional changes were made to the thermal cycling parameters
because of this. First, Foster and Laurin (2012) suggest a room temperature (e.g., 25°C) final
hold as opposed to 4°C to prevent -A. Therefore, a 25°C final hold was evaluated in comparison
to 4°C, and though it did not demonstrate any differences with regard to -A (none were
detected for either data set), I decided to implement the change because using a 25°C final hold
nearly doubled average peak heights from 699rfu to 1203rfu while slightly decreasing total
amplification time. Second, a 13min final extension was evaluated in comparison to the 10min
119
final extension that had been used for the majority of the development process. Using the
longer final extension, -A was completely eliminated.
Completion of the fast PCR development using AmpliTaq Gold Fast with the Identifiler
primer set resulted in a 1hr 19min amplification protocol, which is about one half of the time
needed to perform standard Identifler amplification on the same thermal cycler (see Table 2.6).
Based on the data that utilized the selected protocol (n=18), full profiles were obtained from all
samples (0.25-0.75ng) with an average peak height of 891rfu, an average PHR of 83.6%, 17% of
samples (limited to 0.25ng) exhibiting one or two PHR <50%, an average LPH:TPH CV value of
0.268 and no -A or NSA. Comparison of profile quality obtained from this protocol to those
from standard Identifiler amplification and the other three fast PCR protocols is discussed in the
comparison section below.
3.2.2.3 Identifiler Fast PCR with the SpeedSTAR™ HS DNA Polymerase
The initial evaluation of SpeedSTAR demonstrated its ability to obtain full profiles with
satisfactory peak height and balance, but with a significant amount of non-specific amplfication
and elevated background. Development of a fast PCR protocol using SpeedSTAR and the
Identifiler primer set began with elimination of NSA via an evaluation of primer annealing
temperatures (59°C, 61°C and 63°C) using a 2-step PCR cycle. One sample (a 0.75ng replicate)
from the 59°C data set exhibited a possible low-level NSA peak (178bp at TH01), but no others
were observed with any other sample; this was significantly less than expected based upon the
initial evaluation. However, the initial evaluation involved a sample extracted and quantified
with different chemistries than this study and amplified a greater amount of DNA (1.0ng), but it
120
was unclear if these differences would explain the apparent discrepancy with regard to NSA
peaks. Full profiles were obtained from all samples except for one of the 0.25ng replicates at
59°C. Average peak heights were highest from the 61°C data set (939rfu), followed by 59°C
(904rfu) and 63°C (832rfu), but were acceptable for all data sets. Average PHR was about the
same for all three data sets (77-79%), but was lower than expected (typically >85%). Nearly all
samples exhibited one or more PHR <50% for all temperatures, but instances of PHR <50%
decreased as temperature increased. Inter-locus imbalance was slightly higher for the 59°C and
63°C data sets than the 61°C data set. Even though -A was not observed during the initial
evaluation, I was not surprised to see it for the 0.75ng replicates because the final extension
was only 1min, which had proven to be problematic with the other fast PCR protocols as well.
Furthermore, -A was more frequent for the 63°C data set. Though there was not a temperature
that was clearly superior than the others, 63°C was chosen for further testing because it
exhibited the fewest occurrences of PHR <50%, which proved to be difficult to remedy during
AmpliTaq Gold Fast PCR development.
Initial efforts focused on eliminating the occurrence of PHR <50%. Increasing
denaturation time, as Giese et al. (2009) and Walsh and Erlich et al. (1992) had suggested, was
not a successful approach for AmpliTaq Gold Fast, but switching to 3-step PCR cycling was, as
was using a longer combined annealing/extension step for KAPA2G. Increased denaturation
(from 5sec to 10sec) and annealing/extension (from 25sec to 30sec, 40sec, 50sec and 60sec)
times were evaluated for this product prior to 3-step PCR. Full profiles were obtained from all
samples. As expected, average peak height increased as annealing/extension time increased,
but did not always increase with increased denaturation time. Average PHR for each data set
121
(<81%) was yet again lower than expected (>85%) and numerous loci for all samples exhibited
PHR <50% for all data sets, especially at 0.25ng, but less so using a 5sec denaturation and
40/50/60sec annealing/extension. Inter-locus imbalance decreased as annealing/extension
time increased and was better using 5sec denaturation than 10sec.
At that point, I opted to evaluate 3-step PCR cycling in an attempt to resolve low PHR;
switching to 3-step PCR had been successful using AmpliTaq Gold Fast for this same issue. Three
annealing temperatures (59°C, 61°C and 63°C) and short cycling steps were evaluated and
resulted in allelic dropout, low peak heights, low PHR, inter-locus imbalance and -A for all
annealing temperatures, but no NSA. The 59°C annealing temperature data set had the highest
average peak heights (568rfu, compared to 367rfu and 427rfu for 61°C and 63°C, respectively)
and the least amount of inter-locus imbalance (average LPH:TPH CV of 0.533, compared to
0.732 and 0.598 for 61°C and 63°C, respectively); therefore, it was selected for further testing.
To resolve occurrences of PHR <50%, I evaluated a longer extension time (20sec instead of
10sec) and as a result saw significant improvements in profile completeness, peak height, PHR,
inter-locus balance and -A. However, none of these were at a satisfactory level.
Rather than making the extension step unnecessarily long, I next examined longer
annealing times (15sec compared to 20sec and 25sec) to seek further profile improvements.
This change resulted in full profiles for all samples, with average peak height increasing as
annealing time increased, but peak heights were still too low using 0.25ng (averaging 403rfu)
with the longest annealing time. Occurrences of PHR <50% also significantly declined using a
20sec or 25sec annealing step and were limited to 0.25ng amplifications (except for a single
occurrence at 0.50ng). Despite these improvements, other declines in profile quality were
122
observed: an increase in inter-locus imbalance, -A and definite signs of low-level NSA; all of
these were somewhat expected, given that a longer annealing time allows more amplification
product to form, giving rise to NSA and more -A. Increased -A also reduced inter-locus peak
imbalance because -A did not occur equally at all loci, but tended to occur at TH01 and vWA,
which in turn caused these loci to have lower peak heights because part of the allele was in the
-A form. The 25sec annealing time was selected for further testing despite these issues.
In light of the continued issues with -A, I evaluated longer final extension times of 10min
and 13min, rather than 1min. Using a 10min final extension reduced the occurrence of -A by
93%, but it was not completely eliminated until a 13min final extension was employed. As
predicted by using longer final extensions, inter-locus balance also improved, as did peak
height, so much so that at times peak heights were too high; affected profiles exhibited
increased amounts of called stutter, pull-up and elevated baseline, as well as oversaturation.
PHR <50% continued to occur using 0.25ng and 0.50ng DNA with the longer final extensions. A
variety of low-level NSA peaks was present in all three data sets, but only a 169bp peak at TH01
was consistent observed. Annealing temperatures of 59°C, 61°C and 63°C were once again
evaluated in an attempt to eliminate the formation of NSA products. Full profiles were obtained
from all samples, with similar average peak heights from the three data sets (ranging from
1167rfu to 1333rfu). However, oversaturation was present in the 59°C and 63°C data sets (all
from amplification of 0.75ng DNA), but not the 61°C. NSA was observed in the 59°C data set,
but not 61°C or 63°C. As noted with the previous data sets, more called stutter and pull-up
were observed compared to earlier evaluations, but none exceeded 20% of the true allele. PHR
<50% continued to be problematic for all data sets, occurring in 0.25ng and 0.50ng
123
amplifications, but more so from 0.25ng. Inter-locus peak balance was slightly better from the
61°C data set (0.325) compared to 59°C (0.350) and 63°C (0.412). The 59°C and 61°C data sets
were free of -A, but the 63°C data set was not. Both of these latter two findings had been
previously observed when annealing temperatures were first evaluated for SpeedSTAR. Taking
all of this into account, I selected 61°C for the annealing temperature.
I had attempted to eliminate the occurrence of PHR <50% using 3-step PCR cycling,
longer denaturation, longer annealing (most significant impact) and longer extension, but PHR
>50% could not always be obtained. I thought about increasing annealing time more, but I
suspected that doing so would also result in an increase in peak height. Peak heights were
already on the high end of the acceptable range, and I had previously had problems with peaks
being too high, resulting in oversaturation and increased occurrences of stutter, pull-up and
elevated baseline. I chose not to risk this further and decided not to make any more
modifications to the SpeedSTAR fast PCR protocol.
Completion of the fast PCR development using SpeedSTAR™ HS DNA Polymerase with
the Identifiler primer set resulted in a 49min amplification protocol, which is about one third of
the time needed to perform standard Identifler amplification on the same thermal cycler (see
Table 2.6). Based on the data that utilized the selected protocol (n=18), full profiles were
obtained from all samples (0.25-0.75ng) with an average peak height of 1195rfu, an average
PHR of 83.6%, 6% of samples (limited to 0.25ng) exhibiting a single PHR <50%, an average
LPH:TPH CV value of 0.325 and no -A or NSA. Comparison of profile quality obtained from this
protocol to those from standard Identifiler amplification and the other three fast PCR protocols
is discussed in the comparison section below.
124
3.2.2.4 Identifiler Fast PCR with the Type-it® Microsatellite PCR Kit
During the initial evaluation of Type-it, no allele peaks were generated above threshold.
I presumed this to be because of the short cycling times I initially evaluated, which were not
ideal for this particular kit because it was not developed for fast PCR in spite of the fact that the
polymerase (HotStarTaq® Plus DNA Polymerase) utilized in the kit was a fast enzyme. Based on
the information obtained during fast PCR development for the other three protocols, I decided
to begin with an evaluation of 3-step and 2-step PCR cycling using annealing temperatures of
59°C, longer cycling times for both (30sec/45sec/30sec and 30sec/1min 15sec, respectively) and
a 10min final extension to prevent -A. Full profiles were obtained from all samples, with similar
average peak heights (1280rfu and 1209rfu, respectively) and average PHR (82.7% and 84.0%,
respectively), but PHR <50% occurred in 22% of the 3-step PCR samples (and one of the two
positive control) compared to 6% of the 2-step PCR samples. For AmpliTaq Gold Fast and
SpeedSTAR, more PHR <50% were seen using a 2-step PCR cycle than with a 3-step cycle, thus
the improved performance of Type-it with regard to PHR using a 2-step PCR cycle was
unexpected. Inter-locus peak imbalance was slightly higher for 2-step PCR, with an average
LPH:TPH CV of 0.329 compared to 0.303 for 3-step PCR; despite a slight difference, both are
acceptable. The presence of called -A peaks was not expected since a 10min final extenstion
had been used, but 11% of the 3-step PCR exhibited -A; no -A was present in 2-step PCR
samples. No signs of NSA were observed. Overall, both PCR protocols performed fairly well,
especially in comparison to the initial evaluation. I decided to continue evaluating 2-step PCR
cycling due to its superior performance compared to 3-step PCR with regard to PHR and -A.
125
Since full profiles were obtained with acceptable peak heights, inter- and intra-locus
peak balance and were free of -A and NSA, I focused on reducing amplification time without
comprimising profile quality. Two shorter activation times (1min and 2min) were assessed in
comparison to the 5min activation tested previously, but both resulted in abundant allelic
dropout and reduced peak heights, more so using a 1min activation. Next, I evaluated two
shorter denaturation times (10sec and 20sec) in comparison to the 30sec denaturation that was
tested. Like reduced activation time, a reduction in denaturation also resulted in abundant
allelic dropout and reduced peak heights, more so with using 10sec than 20sec. Lastly, I
assessed two shorter combined annealing/extension steps (55sec and 1min 5sec) to compare to
the 1min 15sec time that had been evaluated. Full profiles were obtained from all data sets
with acceptable peak heights and PHR, but inter-locus imbalance increased as
annealing/extension time decreased. With a 1min 5sec annealing/extension step, average
LPH:TPH CV was 0.365, which is slightly higher than I was aiming for ( 0.350), and with 55sec,
average CV had increased to 0.495, which is unacceptable based upon my evaluation criteria.
Thus, I decided to retain the 1min 15sec annealing/extension step.
Attempts to reduce amplification time without compromising profile quality were
unsuccessful, but this is not unexpected given that Type-it was not designed for fast PCR. In the
end, completion of the fast PCR development using the Type-it Microsatellite PCR Kit with the
Identifiler primer set resulted in a 1hr 14min amplification protocol, which is less than half of
the time needed to perform standard Identifler amplification on the same thermal cycler (see
Table 2.6). Based on the data that utilized the selected protocol (n=18), full profiles were
obtained from all samples (0.25-0.75ng) with an average peak height of 1209rfu, an average
126
PHR of 84.0%, 6% of samples (limited to 0.25ng) exhibiting a single PHR <50%, an average
LPH:TPH CV value of 0.329 and no -A or NSA. Comparison of profile quality obtained from this
protocol to those from standard Identifiler amplification and the other three fast PCR protocols
is discussed in the comparison section below.
3.2.2.5 Comparison of Four Fast PCR Protocols for Identifiler
Each of the four fast PCR protocols I developed for the Identifiler primer set were
compared to each other and standard Identifiler amplification (see Figure D.7) using DNA from
25 individuals (0.6ng DNA input) and dilution series from two of the 25 individuals.
3.2.2.5.1 Determination of the Optimal DNA Input Ranges
For all five data sets, >99% of alleles were detected and 100% full profiles were obtained
when 0.250ng DNA was amplified, except for a single allele below threshold using KAPA2G
with 0.375ng DNA (see Figure 3.16). Allele peak height and balance is summarized in Figure
3.17. As expected, average allele peak height increased as DNA input increased and tended to
be higher using SpeedSTAR and Type-it compared to the other three methods. Reproducibility
of peak height was measured via average coefficient of variation per DNA input, which was
0.350 when 0.250ng DNA was amplified using each of the four fast PCR protocols, but was
often >0.350 using standard Identifiler amplification, indicating less reproducibility using
standard amplification compared to fast. Preferential amplification based on locus size proved
especially time consuming, less robust and not as informative as general inter-locus balance
assessments; average CPALS values only exceeded 0.050 for standard Identifiler amplification
127
using 1.50ng and 3.00ng, as well as with Type-it using 3.00ng. For all other profiles, signs of
preferential amplification based on locus size were either non-existent due to weak R2 values
(<0.3) or signs of preferential amplification of smaller loci did exist, but not above the chosen
threshold (4-fold difference between smallest and largest loci). None of the five methods tested
were able to consistently result in the desired level of general inter-locus balance (CV of
LPH:TPH 0.350), but inter-locus peak balance for each of the fast PCR methods was about as
good as or better than standard PCR when 0.250ng DNA was amplified, but CV increased to
undesirable levels for all methods when 3.00ng was amplified. Intra-locus balance was
measured via average heterozygote peak height ratios, which expectedly increased as DNA
input increased. Instances of PHR <50% tended to occur when 0.188ng DNA was amplified for
any of the five PCR methods and rarely occurred with higher DNA inputs.
Of the various artifacts that were observed above threshold, stutter (n-4) was the most
prevalent (see Figure 3.18). Average percent stutter ranged from 9.7-17% and did not differ
much based on DNA input. Instances of stutter increased as DNA input increased, tending to be
more prevalent using AmpliTaq Gold Fast and SpeedSTAR, but never occurring using standard
PCR and <3.00ng DNA. The lack of called stutter peaks for standard Identifiler amplification was
likley due to the fact that the stutter filters in GeneMapper® ID were specifically set for that
method. Nonetheless, higher percent stutter peaks did form using all of the fast PCR protocols
compared to standard PCR. Unacceptably high stutter peaks (>20%) occurred occassionally, but
were limited to profiles obtained with AmpliTaq Gold Fast or SpeedSTAR. Other forms of stutter
(n+4 and n-8) did occur, but these peaks were not above analysis threshold, except for a single
n-8 stutter peak obtained using SpeedSTAR and 3.00ng DNA (data not shown).
128
Figure 3.16. Sensitivity of standard Identifiler and four fast PCR protocols. Sensitivity is displayed as percent alleles detected and percent full profiles (n=3 per DNA input).
Figure 3.17. Peak height summary for standard Identifiler and four fast PCR protocols. Average peak height, reproducibility of peak height per allele (CV of PH), inter-locus peak balance (CPALS and CV of LPH:TPH) and intra-locus peak balance/imbalance (average PHR and average number of PHR <50% per profile) are displayed for each DNA input (n=3 samples per DNA input).
0%
20%
40%
60%
80%
100%
Average Percent Alleles Detected
0%
20%
40%
60%
80%
100%
Percent Full Profiles
Identifiler AmpliTaq Gold Fast Kapa2G SpeedSTAR Type-it
0
1,000
2,000
3,000
4,000
rfu
Average Peak Height
0.00
0.10
0.20
0.30
0.40
0.50
0.60
CV
Reproducibility: Average CV of Peak Height
-0.20
-0.10
0.00
0.10
0.20
0.30
CP
ALS
Inter-Locus Peak Balance: Preferential Amplifcation of Loci Based on Locus Size
0.000.100.200.300.400.500.600.70
CV
of
LPH
:TP
H
General Inter-Locus Peak Balance
0%
20%
40%
60%
80%
100%
PH
R
Intra-Locus Peak Balance: Average PHR
0.00.51.01.52.02.53.03.5
n
Intra-Locus Peak Imbalance: PHR <50%
Identifiler AmpliTaq Gold Fast Kapa2G SpeedSTAR Type-it
129
Figure 3.18. Artifacts for standard Identifiler and four fast PCR protocols. Average percent stutter (n-4), pull-up and -A are displayed, as well as average number of detected stutter (n-4), pull-up, -A, low-level NSA and elevated baseline per profile. Average number of unacceptable stutter (n-4) and pull-up (i.e., those >20% of the true allele) per profile also displayed (n=3 samples per DNA input).
0%
5%
10%
15%
20%
Average Percent Stutter
0
2
4
6
8
n
Average Instances of Stutter
0.0
0.5
1.0
1.5
n
Average Instances of Unacceptable Stutter
0%
10%
20%
30%
Average Percent Pull-up
0
5
10
15
n
Average Instances of Pull-up
0.0
0.5
1.0
1.5
2.0
2.5n
Average Instances of Unacceptable Pull-up
0%
5%
10%
15%
20%
25%
Average Percent -A
0.0
1.0
2.0
3.0
4.0
5.0
n
Average Instances of -A
0.0
0.5
1.0
1.5
n
Average Instances of NSA
0
2
4
6
n
Average Instances Elevated Baseline
Identifiler AmpliTaq Gold Fast Kapa2G SpeedSTAR Type-it
130
Figure 3.19. Optimal DNA input ranges for standard Identifiler and four fast PCR protocols. An optimal range (red) was determined for each amplification based upon the evaluated criteria (n=3 samples per DNA input).
Sensitivity
Reproducibility
Inter-Locus Balance
Intra-Locus Balance
Stutter
Pull-up
-A
Non-Specific Amp
Baseline
Recommended Range
0.1
25n
g 0
.188
ng
0.2
50n
g 0
.375
ng
0.5
00n
g 0
.750
ng
1.0
0n
g 1
.50
ng
3.0
0n
g
0.1
25n
g
0.1
88n
g 0
.250
ng
0.3
75n
g 0
.500
ng
0.7
50n
g 1
.00
ng
1.5
0n
g 3
.00
ng
0.1
25n
g 0
.188
ng
0.2
50n
g 0
.375
ng
0.5
00n
g 0
.750
ng
1.0
0n
g 1
.50
ng
3.0
0n
g
0.1
25n
g 0
.188
ng
0.2
50n
g 0
.375
ng
0.5
00n
g 0
.750
ng
1.0
0n
g 1
.50
ng
3.0
0n
g
0.1
25n
g
0.1
88n
g 0
.250
ng
0.3
75n
g 0
.500
ng
0.7
50n
g 1
.00
ng
1.5
0n
g 3
.00
ng
Identifiler AmpliTaq Gold Fast KAPA2G SpeedSTAR Type-it
131
Pull-up peaks were the next most abundant artifact that were detected above
threshold, but were nearly always limited to amplification of 1.00ng DNA. Pull-up was more
frequent for SpeedSTAR and Type-it, which also had higher percent pull-up than other
methods. These two methods were also the only ones to generate profiles with unacceptably
high pull-up (percent pull-up >20%), but this was limited to amplification of 3.00ng DNA and
was likely associated with higher average peak heights obtained using SpeedSTAR and Type-it.
Other artifacts included -A, possible low-level non-specific amplification peaks and elevated
baseline, but these were infrequent and tended to be limited to amplification of 3.00ng DNA.
The optimal DNA input range was determined for each of the five amplication protocols
based on the above analyses (see Figure 3.19). The largest optimal range (0.250-1.50ng) was
obtained from standard Identifiler reactions and fast PCR using KAPA2G. Though the fast PCR
protocols using SpeedSTAR and Type-it were superior to the other methods with regard to
sensitivity (full profiles were obtained from 0.125-3.00ng), neither were able to generate
profiles without PHR <50% using 0.250ng DNA.
3.2.2.5.2 Stochastic Threshold
Stochastic thresholds were determined for each fast PCR method in order to identify a
threshold at which heterozygous loci would not be mistaken for homozygous in the event of
allelic drop out; these were compared to that of standard PCR (see Figure 3.20). Fast PCR
stochastic thresholds ranged from 75rfu (KAPA2G) to 165rfu (AmpliTaq Gold Fast), compared to
100rfu for standard PCR. It was not surprising that AmpliTaq Gold Fast most closely resemebled
standard Identifiler, given that AmpliTaq Gold polymerase was used in that amplification.
132
Figure 3.20. Stochastic thresholds for standard Identifiler and four fast PCR protocols. These were established for each amplification method by determing the point at which heterzygous loci would not be mistaken for homozygous loci due to dropout of a single allele (n=447 to 522 loci for each amplification method).
3.2.2.5.3 Precision
Precision of allele sizing was assessed for inter-injection and intra-injection sizing
precision based upon multiple injections of positive control 9947A amplified using each of the
four fast PCR methods and compared to that of standard Identifiler. All assessments indicated
more precise allele sizing between injections (i.e., within a capillary) compared to intra-injection
(across capillaries) precision (see Figure 3.21). Average standard deviations of allele sizing was
within manufacture recommendations (<0.15) for all methods, and observed maxima never
exceeded 0.15 either. The minimum difference between sizing calls for the same allele was
always zero for inter- and intra-injection assessments, regardless of amplification method.
Average and maximum differences were well below 0.50bp for all methods, which is required
for correct allele calling using an allelic ladder. Therefore, all amplification methods exhibited
satisfactory precision for allele sizing.
0
20
40
60
35 50 75 100 125 150 175 200 225 250 275
Loci
wit
h E
xtre
me
Dro
po
ut
Threshold (rfu)
Identifiler AmpliTaq Gold Fast KAPA2G SpeedSTAR Type-it
133
Figure 3.21. Precision of allele sizing for standard Identifiler and four fast PCR protocols. Precision was assessed via standard deviation, as well as base-pair sizing differences, for the same allele within and between capillary electrophoresis injections (n=78 alleles per data set). 3.2.2.5.4 Stutter Assessment
For the detailed stutter analysis, n-4 stutter was the most frequent and accounted for
85% (SpeedSTAR) to 97% (KAPA2G) of all stutter peaks. Percent stutter (n-4) varied by
amplification method (lowest from standard Identifiler and highest from AmpliTaq Gold Fast)
and locus (tended to be lowest at TH01 and TPOX and highest at D21, D3, D2 and D18) (see
Figure 3.22). Use of a global 20% n-4 stutter allowance for all loci is common practice (Myers et
al., 2012) and would work well for any of the developed fast PCR methods. It should be noted,
0.000
0.020
0.040
0.060
0.080A
vera
ge S
tan
dar
d
Dev
iati
on
0.000
0.050
0.100
0.150
Max
imu
m
Stan
dar
d D
evia
tio
n
0.000
0.020
0.040
0.060
Ave
rage
Siz
ing
Dif
fere
nce
(b
p)
0.00
0.05
0.10
0.15
0.20
Standard AmpliTaq Gold Fast KAPA2G SpeedSTAR Type-itMax
imu
m D
iffe
ren
ce
(b
p)
Inter-Injection Intra-Injection
134
however, that the maximum observed percent stutter (n-4) was above 20% for all fast PCR
methods, ranging from 20-26%. These were attributed to a single sample at locus D18; this
sample exhibited an 18% stutter peak when amplified with standard Identifiler.
Both n+4 and n-8 stutter occurred significantly less, often not occurring at many loci
(data not shown). Average n+4 percent stutter ranged from 3% (standard Identifiler, AmpliTaq
Gold Fast and SpeedSTAR) to 5% (KAPA2G and Type-it), whereas average n-8 percent stutter
ranged from 1% (Type-it) to 2% (all other methods).
Figure 3.22. Percent stutter (n-4) for standard Identifiler and four fast PCR protocols. Stutter is displayed by amplification method and locus (n=608 to 792 stutter peaks per method).
3.2.2.5.5 Optimization Check
Twenty-five samples were processed using each of the four fast PCR methods, and the
resulting profiles were compared to those obtained from standard Identifiler amplification. Full
profiles were obtained from 96% of the samples using each of the fast PCR methods, compared
to 92% using standard PCR, with an average of 98% alleles detected using each of the fast PCR
methods and 96% of alleles using standard PCR; all profiles exhibited concordant allele calls and
0%
2%
4%
6%
8%
10%
12%
14%
Standard AmpliTaq Gold Fast KAPA2G SpeedSTAR Type-it
135
no unexplained alleles were present. Inter-locus peak balance was assessed via the average CV
of LPH:TPH and was <0.350 for all methods except fast PCR using Type-it (average 0.380). Intra-
locus peak balance was assessed via average PHR, occurrences of PHR <50% and signs of
greater imbalance due to the magnitude of allele separation and/or locus. All methods
exhibited average PHR between 82.7% and 84.7%, and all methods exhibited at least one
profile with a single PHR <50%. Standard and fast PCR with KAPA2G exhibited 4% of profiles
(i.e., one sample) with a single PHR <50%, AmpliTaq Gold Fast and Type-it each exhibited 8% of
profiles with PHR <50%, while SpeedSTAR exhibited 32% of profiles with PHR <50% for one or
more loci. When examined for increased intra-locus imbalance based on the magnitude of
allele base-pair difference and/or locus, each of the five methods exhibited statistically
significant differences with regard to one or both of these criteria (two-way ANOVA, α=0.05);
however, when the data were further examined for correlation between PHR with either
magnitude of allele difference or locus size, correlation was extremely weak. Therefore, a
decrease in PHR with regard to either an increase in the magnitude of allele repeat difference
or an increase in locus size could not be substantiated for any of the amplification methods.
Lastly, first pass success rates were assessed for each amplification method and ranged from
64% (SpeedSTAR) to 92% (KAPA2G), compared to 88% for standard Identifiler (as well as
AmpliTaq Gold and Type-it); all failing profiles were due to either allelic dropout or PHR <50%.
When all aspects of these analyses were taken into account, KAPA2G™ Fast Multiplex
PCR Kit demonstrated superior performance over the other fast PCR methods – primarily due to
its large optimal DNA input range (0.250-1.50ng), low stochastic threshold (75rfu) and high first
pass success rate (92%) – and was chosen for development of additional fast PCR protocols.
136
3.2.3 Development of Additional Fast PCR Protocols using KAPA2G™ Fast Multiplex PCR Kit
3.2.3.1 Development of 5µl and 6µl Identifiler Fast PCR Protocols
Per manufacturer recommendations, standard Identifiler amplification on 9700 thermal
cycler utilizes the “9600 Emulation” mode, but the 9700 also offers a “Max” mode with faster
ramp rates. A comparison of fast PCR performance using the 9600 emulation and max modes
demonstrated slightly higher quality STR profiles using the maximum ramp rate mode, but
improvements were still needed with respect to lower than desired peak heights (which caused
dropout) and inter-locus peak imbalance. Thus, the number of amplification cycles was
increased from 26 (used by Cellmark for 3µl standard amplifications) to 28 (used by Cellmark
for 5µl standard amplifications), but this introduced -A peaks. Efforts continued with the 26
cycle amplification, coupled with longer annealing/extension times of 45sec and 50sec
(compared to 40sec), but did not result in a significant improvement in allele dropout.
Therefore, a 27 cycle amplification (used by Cellmark for 6µl standard amplifications) was
evaluated next with a 40sec annealing/extension step, which resulted in nearly doubling
average peak height from 503rfu to 976rfu; furthermore, full profiles were obtained from all
samples using 27 cycles, but -A was also present. A 10min final extension successfully
eliminated -A (n=24) and resulted in a 51min amplification protocol.
The 24 samples that were amplified using the final 5µl and 6µl Identifiler fast PCR
protocols exhibited 100% allele concordance to profiles obtained using standard PCR, had a
100% pass rate compared to 83% using standard PCR (all failures due to PHR<50%). It should be
noted that standard PCR pass rates from these 24 samples were lower than normal (typically
>90%) because samples were specifically selected that had PHR <50% using standard PCR to
137
better assess fast PCR’s intra-locus balance. Peak height (averages of 1153rfu and 1205rfu,
respectively), intra-locus (average PHR of 87.4% and 88.8%, respectively) and inter-locus peak
balance (average CV of LPH:TPH of 0.323 and 0.344) were all acceptable from 5µl and 6µl fast
amplifications. These values could not be directly compared to standard PCR because the
original standard amplification was a 6µl reaction detected using POP-6/22cm array.
3.2.3.2 Development of 3µl Identifiler Plus Fast PCR Protocol
Optimization of the 3µl Identifiler Plus PCR protocol began with assessing different final
extension lengths to prevent the formation of -A. Similar to the other fast PCR protocols, -A was
present using the 1min and 5min final extensions, but was eliminated using 10min. However,
low-level NSA was observed using the 59°C annealing/extension temperature, primarily at TH01
(~169bp, ~185bp and occasionally at ~187bp), D16 (~287bp), vWA (~153bp) and TPOX
(~220bp). Therefore, 61°C and 63°C were evaluated to improve primer specificity; NSA
continued at a lesser extent with the use of 61°C, but was eliminated using 63°C. Since full
profiles were obtained from all samples, average peak height (1015rfu) and inter-locus peak
balance (average CV of LPH:TPH of 0.292) were acceptable and all PHR were >50% using 63°C,
that temperature was selected, and development shifted to reducing amplification time.
Shorter initial activation times of 1min and 2min were assessed in comparison to 3min,
both of which exhibited acceptable profiles, though average peak height was reduced to 874rfu
and 849rfu, respectively, while average CV of LPH:TPH improved somewhat (0.278 and 0.275,
respectively) for both activation times. Therefore, the 1min activation was selected for further
testing. Shorter denaturation times (5sec and 10sec, compared to 15sec) were assessed next.
138
Both data sets yielded full profiles, but the 5sec data set exhibited a decrease in average peak
height (711rfu), while the 10sec data set exhibited an increase (1269rfu), compared to 15sec.
Though the explanation for this is unknown, I decided to continue with a 10sec denaturation
because it performed as well or better than 15sec with regard to peak height and other profile
quality criteria, whereas average peak height using 5sec was getting too low.
Therefore, I next evaluated a 10sec denaturation with a 50sec annealing/extension step
in comparison to the 60sec annealing/extension step previously tested, using a large data set
(n=88) for both in an effort to reduce skewing due to small sample size; fast profiles from these
88 samples were compared to those obtained using standard PCR. Full profiles were obtained
from 93%, 91% and 94%, respectively, with an average of 96% of alleles detected from both of
the fast protocols compared to 98% from standard. Fast PCR’s slightly increased rate of dropout
and decreased profile completeness were likely a result of reduced peak heights obtained from
fast PCR (967rfu for the 50sec data set and 675rfu for 60sec) compared to standard PCR
(1303rfu). It should be noted however, that fast PCR was performed on a different thermal
cycler and detected on a different 3130xl Genetic Analyzer than the standard PCR data, which
could account for the differences seen in peak height. Inter-locus peak balance was acceptable
from all methods, with average CVs of LPH:TPH ranging from 0.247 for standard PCR to 0.258
(50sec) and 0.261 (60sec) fast PCR protocols, which is considerably lower than that observed by
Identifiler fast PCR (typically >0.300). PHR <50% were infrequent, occurring in 1% of fast PCR
profiles (50sec data set only; none observed in the 60sec data set) compared to 3% of standard
profiles. Overall, first pass success rates were highest from the 50sec fast PCR data set (92%),
compared to 91% for 60sec fast PCR and standard PCR (dropout and PHR <50% were the only
139
reasons for sample failure). Thus, the 50sec annealing/extension step was selected for the
optimized Identifiler Plus fast PCR protocol, which had a total run time of 49min.
3.2.3.3 Development of 3µl PowerPlex 16 HS Fast PCR Protocol
Optimization of the 3µl PowerPlex 16 HS PCR protocol also began with assessing
different final extension lengths to prevent the formation of -A. Similar to the other fast PCR
protocols, -A was present using the 1min and 5min final extensions, but was eliminated using
10min. Furthermore, from the first round of testing with the 10min final extension step, full
profiles were obtained from all samples, no signs of NSA were present, average peak heights
were 1256rfu, PHR <50% (17% of samples) were limited to 0.25ng samples, but inter-locus
imbalance was higher than desired (average CV of LPH:TPH was 0.438). Next, 2-step PCR cycling
was evaluated using annealing/extension temperatures of 58°C, 60°C and 62°C in an effort to
improve profile quality. Low-level NSA was observed using both of the lower temperatures, but
not at 62°C; however, profile quality decreased using 2-step PCR, most notably via allelic
dropout (11% of samples) despite increased average peak heights (1476rfu), increased
occurrences of PHR <50% (28% of samples) and decreased inter-locus balance (average CV of
LPH:TPH of 0.474). Therefore, 2-step was not pursued further, and development of 3-step
cycling continued with an assessment of increasing ramp rates from 29% and 23% to 100% for
the annealing and extensions, respectively. Compared to the use of manufacturer’s
recommended ramp rates, use of 100% ramp rates exhibited allelic dropout from one sample (a
0.25ng replicate; 6% of samples), despite increased average peak heights (1580rfu), decreased
occurrences of PHR <50% (11% of samples; also limited to 0.25ng) and similar inter-locus
140
balance (average CV of LPH:TPH of 0.446); however, increased average instances of called
stutter per sample and pull-up >20% accompanied these higher peak heights. Use of 100%
ramp rates was further tested in conjunction with reduced annealing times (15sec compared to
30sec). Even though full profiles were obtained from all samples using the shorter annealing
time, a significant reduction in average peak height was observed (890rfu) and occurrences of
PHR <50% doubled (22% of samples; limited to 0.25ng), while occurrences of stutter and pull-
up declined; inter-locus peak balance was not significantly effected (average CV of LPH:TPH of
0.422). Since peak heights were still at an acceptable level using 100% ramp rates and 15sec
annealing, these were assessed next with shorter extension times (20sec and 15sec, compared
to 30sec). Extension times less than 30sec resulted in a significant increase in allelic dropout
and reduction in peak heights, such that full profiles were only obtained from 73% (15sec) and
82% (20sec) of samples. Thus, the 30sec extension was maintained. Reduced denaturation time
(10sec and 5sec, compared to 15sec) was assessed next, but as seen with a reduction in
extension time, a reduction in denaturation time also resulted in significant increases in allelic
dropout and decreases in peak height. Therefore, 15sec denaturation was maintained. Next, I
assessed a reduction in initial activation (1min versus 2min) and found that full profiles were
still obtained from all samples, while average peak height (914rfu), inter-locus balance (average
CV of LPH:TPH of 0.414) and occurrences of PHR <50% (no occurrences) were not effected in a
negative manner. The 1min initial activation was then tested with a 25°C final hold in
comparison to 4°C, from which full profiles were obtained, a significant increase in average
peak height was observed (1888rfu), accompanied by increased occurrences of pull-up (not
exceeding 20% of the true allele), but not significant changes in inter- or intra-locus balance.
141
The implementation of the 25°C final hold completed the development of the PowerPlex 16 HS
fast PCR protocol, resulting in a 51min amplification protocol. It should be noted that the
desired level of inter-locus peak balance (CV of LPH:TPH 0.350) could not be achieved.
This fast PCR protocol was tested using 89 samples, and the resulting profiles were
compared to those obtained using standard PCR. Full profiles were obtained from 96% of fast
profiles and 93% of standard profiles, with an average of 98% and 99% of alleles detected,
respectively. Peak heights were slightly lower from fast PCR (average of 1308rfu) compared to
standard (average of 1449rfu), but were still on the high end of the desired range. Inter-locus
peak imbalance was higher than desired using fast and standard PCR (average CVs of LPH:TPH
of 0.416 and 0.372, respectively). PHR <50% were infrequent, occurring in 2% of fast PCR
profiles compared to 4% of standard profiles. Overall, first pass success rates were highest from
fast PCR (94%), compared to 90% for standard PCR; dropout and PHR <50%, were the only
reasons for sample failure. See Figure D.10 for a representative fast PCR profile.
3.2.4 Validation of Fast PCR Protocols using KAPA2G™ Fast Multiplex PCR Kit
3.2.4.1 Determination of the Optimal Range of Input DNA
For all five fast PCR protocols (see Figures D.8 – D.10 for electropherograms), average
percent alleles detected and percent full profiles increased as template amount increased, such
that full profiles were obtained from nearly all samples when 0.188ng DNA was amplified (see
Figure 3.23). Though the senstivity range (i.e., the range in which full profiles were obtained
from the majority of all samples) was determined to be 0.188-3.00ng for all but 3µl Identifiler
fast PCR (0.375-3.00ng), data from 0.188-3.00ng is discussed further for each of the five fast
142
PCR methods for comparison purposes. Allele peak height and balance is summarized in Figure
3.24. As expected, average allele peak height increased as DNA input increased. Reproducibility
of peak height was measured via average coefficient of variation per DNA input, which was
0.350 when 0.188ng DNA was amplified using each of the four fast PCR protocols (except for
0.188ng with a 6µl Identifiler amplification), indicating acceptable levels of reproducibility.
None of the five methods tested were able to result in the desired level of general inter-
locus balance (CV of LPH:TPH 0.350) for all DNA input amounts. Nearly all methods exhibited
CVs >0.350 at 0.188ng and 3.00ng, while PowerPlex 16 HS CVs were >0.350 for all but 3.00ng.
Intra-locus balance was measured via heterozygote peak height ratios, averaging >82% for all
0.375ng amplifications. Instances of PHR <50% were most frequent when 0.188ng DNA was
amplified, but on average occurred less than once per sample.
Figure 3.23. Sensitivity of fast PCR protocols using KAPA2G. Sensitivity is displayed as percent alleles detected and percent full profiles (n=5 or 6 per DNA input; those with n=5 had a sample removed due to injection failure).
0%
20%
40%
60%
80%
100%
Average % Alleles Detected
0%
20%
40%
60%
80%
100%
Percent Full Profiles
Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex16 HS (3µl)
143
Figure 3.24. Peak height summary for fast PCR protocols using KAPA2G. Average peak height, reproducibility of peak height per allele, inter-locus peak balance and intra-locus peak balance/imbalance are displayed for each DNA input (n=5 or 6 per DNA input; those with n=5 had a sample removed due to injection failure). Various artifacts were observed above the analysis threshold, but stutter (n-4) was the
most prevalent type (see Figure 3.25). Average percent stutter ranged from 10-19% and
demonstrated a slight increase for the 0.375ng samples compared to 0.750-3.00ng. This was
expected given that peak heights were lower at 0.375ng compared to higher inputs; thus, any
stutter peaks that met the 75rfu analysis threshold at 0.375ng were a larger percentage of the
true allele peak. Instances of stutter increased as DNA input increased and tended to be much
0
1,000
2,000
3,000
4,000
5,000
0.188ng 0.375ng 0.750ng 1.50ng 3.00ng
rfu
Average Peak Height
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.188ng 0.375ng 0.750ng 1.50ng 3.00ng
CV
Reproducibility: Average CV of Peak Height
0.00
0.20
0.40
0.60
0.80
0.188ng 0.375ng 0.750ng 1.50ng 3.00ng
CV
of
LPH
:TP
H
General Inter-Locus Peak Balance
0%
20%
40%
60%
80%
100%
0.188ng 0.375ng 0.750ng 1.50ng 3.00ng
PH
R
Intra-Locus Peak Balance: Average PHR
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
0.188ng 0.375ng 0.750ng 1.50ng 3.00ng
n
Intra-Locus Peak Imbalance: PHR <50%
Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex16 HS (3µl)
144
less prevalent from PowerPlex 16 HS than the Identifiler/Identifiler Plus profiles. Unacceptably
high stutter peaks (>20%) occurred occassionally, but were limited to profiles obtained using
3.00ng DNA (data not shown). Other forms of stutter (n+4 and n-8) did occur above threshold
on occassion, but were nearly always limited to 1.50ng and 3.00ng amplifications. Furthermore,
average percent stutter for these two forms of stutter (<4% for all) was much lower than that of
n-4 stutter. Unacceptably high n-8 stutter (>2 occurrences in a single profile) occurred was
limited to a single 3.00ng Identifiler Plus amplification (data not shown).
Figure 3.25. Artifacts for fast PCR protocols using KAPA2G. Average percent stutter and pull-up are displayed, as well as average number of detected stutter, pull-up and elevated baseline per profile (n=5 or 6 samples per DNA input; those with n=5 had a sample removed due to injection failure).
0%
5%
10%
15%
20%
Average Percent Stutter (n-4)
0
2
4
6
8
n
Average Instances of Stutter (n-4)
0%
1%
2%
3%
4%
Average Percent Stutter (n+4)
0
1
1
2
2
n
Average Instances of Stutter (n+4)
0%
2%
4%
6%
8%
10%
Average Percent Stutter (n-8)
0
1
2
3
4
n
Average Instances of Stutter (n-8)
0%
2%
4%
6%
8%
10%
Average Percent Pull-up
0
5
10
15
20
25
n
Average Instances of Pull-up
0
2
4
6
n
Average Instances of Elevated Baseline
Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex16 HS (3µl)
145
Pull-up peaks were the next most abundant artifact that was detected above threhsold,
but were nearly always limited to amplification of 1.50ng DNA and were more frequent for
Identifiler Plus and PowerPlex 16 HS. Unacceptably high pull-up (>20%) was limited to 3.00ng
amplifications using Identifiler (6µl) and Identifiler Plus (data not shown). A single occurrence of
-A (2.7% of the true allele) was present in a 3µl Identifiler amplification using 3.00ng DNA (data
not shown). Elevated baseline was limited to 1.50ng and 3.00ng and was unacceptably high
(occurrences at >3 loci) at 3.00ng for all five PCR methods (data not shown). No signs of non-
specific amplification were noted.
The optimal DNA input range was determined to be 0.375-1.50ng for each of the five
fast PCR protocols based on the above analyses (see Figure 3.26).
3.2.4.2 Stochastic Threshold
Stochastic thresholds were determined for each fast PCR method and were compared to
those of standard PCR (see Figure 3.27). Fast PCR stochastic were as good as or better (85-
125rfu) than those from standard PCR (120-270rfu).
146
Figure 3.26. Optimal DNA input ranges for fast PCR protocols using KAPA2G. An optimal range (red) was determined for each amplification based upon the evaluated criteria (n=5 or 6 samples per DNA input; those with n=5 had a sample removed due to injection failure).
Sensitivity
Reproducibility
Inter-Locus Balance
Intra-Locus Balance
Stutter
Pull-up
-A
Non-Specific Amp
Baseline
Recommended Range
0.1
88n
g
0.3
75n
g
0.7
50n
g
1.5
0n
g
3.0
0n
g
Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex 16 HS (3µl)
0.1
88n
g
0.3
75n
g
0.7
50n
g
1.5
0n
g
3.0
0n
g
0.1
88n
g
0.3
75n
g
0.7
50n
g
1.5
0n
g
3.0
0n
g
0.1
88n
g
0.3
75n
g
0.7
50n
g
1.5
0n
g
3.0
0n
g
0.1
88n
g
0.3
75n
g
0.7
50n
g
1.5
0n
g
3.0
0n
g
147
Figure 3.27. Stochastic thresholds for fast PCR protocols using KAPA2G. These were established for each amplification method by determing the point at which heterzygous loci will not be mistaken for homozygous loci due to dropout of a single allele (n=510 to 595 loci for each amplification method).
3.2.4.3 Precision
Precision of allele sizing was assessed for inter-injection and intra-injection sizing
precision based upon multiple injections of positive control 9947A amplified using each of the
3μl fast PCR methods that were developed for each of the three primer sets tested. Upon initial
review of the data, the Identifiler and Identifiler Plus data sets suffered from “first injection
effect” and thus, the first injection worth of data was removed only for Identifiler and Identifiler
Plus inter-injection assessments. Average standard deviations of allele sizing were <0.15 for all
methods, whereas observed maximum standard deviation exceeded 0.15 for intra-injection of
Identifler and Identifiler Plus only (see Figure 3.28). The majority of alleles exhibiting standard
deviations >0.15 had allele sizes greater than 250bp (71% for Identifiler and 100% for Identifiler
Plus) and accounted for a small proportion of the alleles tested (9.0% for Identifiler and 2.6%
0
20
40
60
80
100
35 50 75 100 125 150 175 200 225 250 275
Loci
wit
h E
xtre
me
Dro
po
ut
Threshold (rfu)
IDF (3µl) IDF (5µl) IDF (6µl) ID+F (3µl) HSF (3µl)
IDS (3µl) IDS (5µl) IDS (6µl) ID+S (3µl) HSS (3µl)
148
for Identifiler Plus). Furthermore, average and observed maximum allele sizing differences were
<0.50bp for all methods (see Figure 3.28).
As was seen previously with the comparison of various Identifiler fast PCR protocols,
nearly all assessments indicated more precise allele sizing between injections (using the
corrected data sets) compared to intra-injection precision. Furthermore, all amplification
methods exhibited satisfactory precision for allele sizing.
Figure 3.28. Precision of allele sizing for fast PCR protocols using KAPA2G. Precision was assessed via standard deviation, as well as base-pair sizing differences, for the same allele within and between capillary electrophoresis injections (n=75 alleles for PowerPlex 16 HS and 78 for Identifiler and Identifiler Plus data sets).
0.000
0.050
0.100
Ave
rage
Sta
nd
ard
D
evia
tio
n
0.000
0.100
0.200
0.300
Max
imu
m
Stan
dar
d D
evia
tio
n
0.000
0.050
0.100
Ave
rage
Siz
ing
Dif
fere
nce
(b
p)
0.000
0.100
0.200
0.300
Identifiler Identifiler Plus PowerPlex 16 HS
Max
imu
m S
izin
g D
iffe
ren
ce (
bp
)
Inter-Injection Intra-Injection
149
3.2.4.4 Stutter Assessment
For the detailed stutter analysis, n-4 stutter was the most frequent type of stutter
observed and accounted for 82% (3µl PowerPlex 16 HS) to 96% (6µl Identifiler) of all stutter
peaks. Percent stutter (n-4) varied more by locus than amplification method (see Figure 3.29).
Maximum observed stutter for each fast PCR method exceeded the locus specific stutter
thresholds supplied by vendors for each primer set processed under their recommendations
(i.e., standard PCR, using 25μl reaction volumes)(see Figure 3.30). Unless otherwise modified by
the user in the GeneMapper® ID Panel Manager, the vendor-specific thresholds will be used by
GeneMapper® ID during analysis, and due to the increase in percent stutter for fast PCR, more
stutter peaks will be called using these fast PCR methods compared to standard. Thus,
laboratories desiring to implement fast PCR (or other amplification methods with stutter
thresholds different than those supplied by the vendor) should be aware of this issue.
Figure 3.29. Average percent stutter (n-4) for fast PCR protocols using KAPA2G. Stutter is displayed by amplification method and locus (n=1592 to 2044 stutter peaks per method). D2S1338 and D19S433 are Identifiler and Identifiler Plus loci, while Penta D and Penta E are PowerPlex 16 HS loci.
0%
2%
4%
6%
8%
10%
12%
14%
Per
cen
t St
utt
er
IDF (3µl) IDF (5µl) IDF (6µl) ID+F (3µl) HSF (3µl)
150
As mentioned previously, implementing global stutter thresholds of 20% for all loci is
not an uncommon practice. Maximum stutter limits per locus were calculated via the sum of
average percent stutter plus three standard deviations (a more conservative approach adds two
standard deviations; Verheij et al., 2012) and never exceeded 20% for any locus or amplification
method. It should be noted, however, that the maximum observed percent stutter (n-4) was
occasionally (0.05-0.18% of samples) above 20% for all Identifiler and Identifiler Plus fast PCR
methods, ranging from 21-25%, but never exceeded 20% for PowerPlex 16 HS. Half of the
profiles exhibiting n-4 stutter >20% had stutter peaks that corresponded with pull-up from
another locus, 33% were from allele 25 at D18 and the remaining 17% were from loci with low-
level peak heights (<140rfu), which are subject to stochastic effects. Thus, a global 20% n-4
stutter threshold would work well for any of the developed fast PCR methods.
Both n+4 and n-8 stutter occurred significantly less than n-4, often not occurring at loci
(data not shown). Since occurrences of n+4 and n-8 were often low for individual loci, these
types of stutter were averaged for loci. Average n+4 percent stutter ranged from 2% to 4%,
whereas average n-8 percent stutter ranged from 2% to 3% (see Table 3.6). Maximum allowable
n+4 and n-8 stutter were calculated for all loci as opposed to individual loci; however,
maximum observed n+4 exceeded calculated maximums for three fast PCR protocols (5µl
Identifiler, 3µl Identifiler Plus and 3µl PowerPlex 16 HS), while maximum observed n-8
exceeded calculated maximums for all five protocols. These discrepancies could arise from the
fact that all stutter peaks of the same type were grouped together because not enough data
existed to calculate these values for individual loci. Given that many of Cellmark’s databasing
clients allow a 20% n+4 stutter threshold, and the maximum observed n+4 stutter was 16.6%, it
151
appeared reasonable to apply the same 20% global stutter threshold to n+4 stutter for all five
fast PCR protocols. Most databasing clients don’t provide specific requirements for n-8 stutter;
therefore, based upon the data obtained, a 12% global stutter threshold for n-8 stutter was
applied to the Identifiler fast PCR protocols and 10% for Identifiler Plus/PowerPlex 16 HS.
Figure 3.30. Maximum observed stutter (n-4) for fast PCR protocols using KAPA2G compared to standard PCR. Maximum observed stutter was nearly always higher than the vendor-specified, locus specific stutter thresholds for standard, full volume PCR reactions of each primer set.
0%5%
10%15%20%25%30%
Per
cen
t St
utt
er
Identifiler
0%
5%
10%
15%
20%
25%
Per
cen
t St
utt
er
Identifiler Plus
0%
5%
10%
15%
20%
25%
Per
cen
t St
utt
er
PowerPlex 16 HS
IDF (3µl) IDF (5µl) IDF (6µl) ID+F (3µl)
HSF (3µl) ID ID+ HS
152
Table 3.6
Average Percent Stutter for n+4 and n-8 Stutter
Stutter Type Identifiler
(3µl) Identifiler
(5µl) Identifiler
(6µl) Identifiler Plus (3µl)
PowerPlex 16 HS (3µl)
n+4 Average 3.82%a 2.97%a,b 3.17%a,b 2.43%b 2.75%b 95% CI [3.09%, 4.56%] [2.53%, 3.40%] [2.26%, 4.08%] [2.09%, 2.78%] [2.45%, 3.05%] Calculated Maximum 10.6% 8.62% 10.4% 7.25% 9.17% Observed Maximum 9.01% 12.2% 8.13% 9.49% 16.6%
n-8 Average 2.52%c,d 2.05%d,e 2.24%c,d,e 1.77%e 2.60%c 95% CI [2.01%, 3.03%] [1.80%, 2.30%] [1.55%, 2.93%] [1.58%, 1.97%] [2.38%, 2.82%] Calculated Maximum 8.52% 6.18% 8.99% 4.98% 7.65% Observed Maximum 10.9% 10.9% 10.4% 7.55% 8.88%
Note. Average percent stutter (n+4 and n-8) is displayed for each fast PCR method. Averages in each row that share subscripts are not statistically different at α=0.05 according to the Tukey HSD procedure. Not enough data existed to calculate average percent stutter by locus for these two types of stutter.
3.2.4.5 Automation (Large Sample Sets)
Each of the five fast PCR protocols were tested using 86-89 buccal samples (swab
cuttings or Buccal DNA Collector™ punches) with known profiles (see Table 3.7 for a summary
of profile quality). Full profiles were obtained from 97% of samples for each fast PCR method,
with an average of 98% alleles detected; all profiles exhibited concordant allele calls and no
explained alleles were present. Inter-locus peak balance was assessed via the average CV of
LPH:TPH and was <0.350 for all fast methods except with the PowerPlex 16 HS primer set, but it
should be noted that inter-locus peak balance is lower than desired using standard PowerPlex
16 HS (3µl amplification; data not shown) as well. Intra-locus peak balance was assessed via
average PHR and occurrences of PHR <50%. All methods exhibited average PHR between 85.8%
and 86.8%, and all but 6µl Identifiler exhibited at least one profile with a single PHR <50%.
Lastly, first pass success rates were 95% for each of the five fast PCR protocols. It should be
noted that all samples exhibiting dropout with a fast PCR method also exhibited dropout when
153
amplified using standard PCR (except for one sample amplified using 6µl Identifiler fast PCR).
On the other hand, PHR <50% were generally not reproducible across PCR methods.
Furthermore, it should be noted that a low level (~100rfu), unexplained artifact was
noted at Amelogenin (~108bp) for 1.1% of 3µl Identifiler and 5µl Identifiler fast amplifications (a
single sample for both amplification volumes; see Figure 3.31). This artifact was not
reproducible and at no other time during fast PCR development or validation was it observed.
Table 3.7
Profile Summary From Automation Assessment
Profile Assessment Identifiler
(3µl)a
Identifiler (5µl)
b
Identifiler (6µl)
c
Identifiler Plus (3µl)
c
PowerPlex 16 HS (3µl)
d
Full Profiles Detected 99% 99% 98% 97% 98% Alleles Detected
Average
98.9%
99.0%
98.1%
98.0%
99.4% 95% CI [96.7%, 100%] [97.0%, 100%] [95.4%, 100%] [95.5%, 100%] [98.4%, 100%]
Peak Height (rfu) Average 827 961 579 1106 1295 95% CI [811, 843] [943, 979] [567, 591] [1076, 1136] [1263, 1326] Observed Maximum 3418 3897 2757 7955 7162 Observed Minimum 119 81 78 76 75
Inter-Locus Balance: CV of LPH:TPH
Average 95% CI
0.294 [0.282, 0.306]
0.301 [0.291, 0.311]
0.306 [0.295, 0.317]
0.264 [0.252, 0.276]
0.416 [0.408, 0.424]
Intra-Locus Balance: PHR Average 95% CI PHR <50%
e
0.866
[0.861, 0.872] 0.9%
0.864
[0.858, 0.870] 1.8%
0.861
[0.855, 0.867] 0.0%
0.868
[0.862, 0.874] 0.9%
0.858
[0.852, 0.864] 2.7%
First Pass Success Rate 98% 97% 98% 95% 95% Failure Reasons
Dropout PHR <50%
f
1.2% 1.2%
1.1% 2.2%
2.3% 0.0%
3.4% 1.1%
2.3% 3.4%
Note. Profile summaries are displayed for each of the five fast PCR protocols. an=86.
bn=89.
cn=87.
dn=88.
ePercent of all heterozygous loci.
fPercent of samples.
154
Figure 3.31. Fast PCR artifact at Amelogenin. A low-level artifact (~108bp) was identified in one 3µl Identifiler amplification (top) and one 5µl Identifiler amplification (bottom). These fast amplifications were of different samples.
3.2.4.6 Contamination Assessment
All negative amplification controls were free of contamination, and no signs of
contamincation due to the amplification process were identified.
3.2.4.7 Lot-to-Lot Variation
Lot-to-lot variation of KAPA2G™ Fast Multiplex PCR Kit was assessed for three different
lots of the master mix (see Table 3.8 for a summary of profile quality). Full profiles were
obtained from all samples and positive amplification controls using each lot. Average peak
heights were significantly higher using Lots 2 and 3, whereas no significant difference in intra-
locus (average PHR and instances of PHR <50%) or inter-locus (average CV of LPH:TPH) balance
155
were observed for the three lots. Despite lower peak heights using Lot 1, high quality, full
profiles were obtained from all three lots.
Table 3.8
Lot-to-Lot Variation of KAPA2G™ Fast Multiplex PCR Kit
Profile Assessment Lot 1 Lot 2 Lot 3
Peak Height (rfu) Average 577 787a 852a 95% CI [534, 620] [725, 849] [793, 911] Observed Maximum 1892 2936 2748 Observed Minimum 138 129 215
Inter-Locus Balance: CV of LPH:TPH Average 95% CI
0.379b
[0.320, 0.438]
0.370b
[0.338, 0.402]
0.331b
[0.289, 0.373] Intra-Locus Balance: PHR
Average 95% CI PHR <50%
0.854c
[0.828, 0.880] 3.8%d
0.857c
[0.836, 0.878] 1.3%d
0.851c
[0.829, 0.873] 0.0%d
Note. Profile summaries are displayed for each of the five fast PCR protocols. Full profiles were obtained from all samples and positive amplification controls. Averages in each row that share subscripts are not statistically different at α=0.05 according to one-way ANOVA analysis and/or the Tukey HSD procedure.
3.2.4.8 Storage Conditions
Storage conditions were assessed via 12 combinations of number of thaws and length of
4°C storage prior to use. Full profiles were obtained for all samples and positive amplification
controls. Peak heights were significantly different for many of the storage conditions (see Table
3.9), but general trends with regard to number of thaws and length of 4°C storage were not
substantiated for the majority of data subsets (see Figure 3.32). Peak height variation due to
different amplification and detection runs on the same instrument is a widely accepted
concept; however, in this particular case, it was unknown whether this additional variation in
peak height influenced the data such that apparent trends appeared to exist for the 3-thaw and
156
immediate amp data sets, when in fact such trends did not exist, or if additional trends existed
but appeared not to. Either way, peak height values were acceptable from all data sets. Intra-
and inter-locus peak balance were assessed via PHR and CV of LPH:TPH, respectively, and did
not exhibit significant differences between storage conditions. Therefore, all storage conditions
tested were suitable for use.
Table 3.9
Variation Due to Storage Conditions of KAPA2G™ Fast Multiplex PCR Kit
Length of 4°C Storage 1-Thaw 2-Thaw 3-Thaw
Peak Heighta Immediate Use One Week Two Weeks One Month
780b,c [745, 815] 671b [645, 697] 837c,d [795, 879] 837c,d [811, 863]
959d,e [919, 999]
1068e [1023, 1113] 630b [606, 654] 891d [833, 949]
1138e [1075, 1201]
871c,d [838, 904] 832c,d [812, 852] 732b,c [710, 754]
PHRf Immediate Use One Week Two Weeks One Month
0.856 [0.850, 0.861] 0.858 [0.851, 0.865] 0.859 [0.853, 0.864] 0.854 [0.848, 0.860]
0.868 [0.865, 0.870] 0.849 [0.839, 0.860] 0.851 [0.845, 0.856] 0.854 [0.844, 0.863]
0.832 [0.830, 0.840] 0.840 [0.829, 0.850] 0.868 [0.863, 0.873] 0.863 [0.853, 0.869]
CV of LPH:TPHg Immediate Use One Week Two Weeks One Month
0.370 [0.338, 0.402] 0.350 [0.302, 0.398] 0.356 [0.305, 0.406] 0.347 [0.295, 0.400]
0.319 [0.268, 0.369] 0.357 [0.316, 0.398] 0.359 [0.302, 0.416] 0.354 [0.299, 0.410]
0.383 [0.333, 0.457] 0.370 [0.329, 0.428] 0.388 [0.332, 0.430] 0.385 [0.329, 0.424]
Note. Peak height, PHR and CV of LPH:TPH averages and 95% confidence intervals are displayed for each storage condition. Peak height averages that share subscripts are not statistically different at α=0.05 according to the Tukey HSD procedure; intra- and inter-locus balance (PHR and CV of LPH:TPH) were not significantly different for any of the storage conditions using two-way ANOVA. an=192.
fn=80.
gn=7.
157
Figure 3.32. Effect of storage conditions on peak height. Average peak height is displayed for the various storage conditions tested (n=7 samples per data set). Consisent trends with regard to peak height were only noted for the 3 Thaws data set (decrease as length of 4°C increased) and the Immediate Amp data set (unexpected increase as number of thaws increased).
3.2.5 Post-Validation Modifications to Identifiler Fast PCR Protocols
As testing progressed with fast PCR amplifications, low level non-specific amplification
was routinely seen at TH01 (~185bp) using the 3µl Identifiler fast amplification. Use of a 63°C
annealing/extension temperature (an increase from the validated 61°C) was tested on 86
samples and demonstrated the elimination of this artifact while still generating high quality
profiles. Full profiles were obtained from 100% of the tested samples, with average peak
heights of 1219rfu, average PHR of 87.7%, no instances of PHR <50% and average CV of
LPH:TPH of 0.315. First pass success rate was 99%, with a single sample failing due to elevated
n-4 stutter and elevated baseline. Representative profiles using 61°C and 63°C annealing
temperatures are displayed in Figure 3.33.
0
500
1000
1500
1 Thaw 2 Thaws 3 Thaws
rfu
Immediate Amp 1 week @ 4°C 2 weeks @ 4°C 1 month @ 4°C
0
500
1000
1500
Immediate Amp 1 week @ 4°C 2 weeks @ 4°C 1 month @ 4°C
rfu
1 Thaw 2 Thaws 3 Thaws
158
A) 61°C Annealing/Extension
B) 63°C Annealing/Extension
Figure 3.33. Effect of annealing/extension temperature on non-specific amplification. Using a 61°C annealing/extension temperature, low-level NSA is present at TH01 in this example (A), but is eliminated when a 63°C annealing/extension temperature is utilized (B).
Additional testing of the 6µl Identifiler fast amplification protocol revealed more
extensive low-level non-specific amplification than was seen for the 3µl amplifications and
included TH01 and TPOX loci; NSA was seen at D13 less often. Increased occurrences of -A in
ChargeSwitch samples and positive amplification controls (9947A DNA) were also observed. The
159
presence of NSA (samples only) was not all that surprising, but the formation of -A in samples
and controls was puzzling given that nothing had changed since development and validation
with regard to how positive controls were amplified. Eventually, thermal cycler and/or KAPA2G
performance came into question, but neither could ever be isolated as the cause of -A.
Longer final extensions with the 6µl amplification were tested first and exhibited
expected and unexpected results. Overall, it was expected that the occurrence of -A peaks
would decrease as final extension time increased. However, as final extension increased from
10min or 13min to 15min, -A occurrence unexpectedly increased, and it wasn’t until final
extension reached 20min that -A decreased (see Figure 3.34). This is likely explained by the fact
that many -A peaks were present but undetected (either because they were below threshold
and/or the -A peak was not resolved enough from the true allele peak for the software to
detect it), thus as peak heights increased with the longer final extension times, so did peak
heights of -A peaks, such that they were detected by the software. Since -A was not completely
eliminated using a 20min final extension, 23min was also tested, but resulted in over-
adenylation (+A). Furthermore, these longer extension times also resulted in increased peak
heights, inter-locus peak imbalance, low-level NSA and n-8 stutter. Though no final extension
time resulted in completely ideal profiles, a 20min final extension was selected because it
resulted in the least disturbance to profile quality. It should be noted that the vast majority of
the samples tested had >1ng of DNA amplified, and occurrences of -A, +A and n-8 stutter were
associated with higher DNA inputs (see Figure 3.35).
The need for a higher annealing temperature became more apparent following the
increase in low-level, non-specific amplification associated with the 20min final extension.
160
Figure 3.34. Effect of final extension length on -A and +A. The percent of samples exhibiting one or more occurrences of -A or +A is displayed by the final extension length (n=10 for 10-20min and 74 for 23min).
Figure 3.35. Effects of final extension length and input DNA on profile quality. Stutter (n-8), -A and +A are strongly associated with high input DNA (n=74).
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
10min 13min 15min 20min 23min
Final Extension Length
-A +A
0 1 2
Input DNA(ng)
6µl ID Fast PCR (20min Final Extension)
0 1 2
Input DNA (ng)
6µl ID Fast PCR (23min Final Extension)
Pass
Dropout
PHR <50%
Elevated Baseline
Stutter (n-8)
Stutter (n+4)
Non-Specific Amplification
+A
-A
161
A) 61°C Annealing/Extension
B) 63°C Annealing/Extension
Figure 3.36. Effects of annealing/extension temperature and final extension on profile quality. The top profile (A) was obtained from the early developmental stages of a proposed first pass procedure for buccal punch samples (2μl beads, 250μl washes, 120μl elution volume, 0.7μl DNA amplified) and underwent a 6μl Identifiler fast PCR amplification per original validation conditions with subsequent detection with POP-4/36cm array using a 3kV 10sec injection. Low-level, non-specific amplification products (at TH01 and TPOX) and -A (at vWA) are circled in red. The bottom profile (B) was obtained under slightly different conditions (0.5μl beads, 125μl washes, 60μl elution volume, 1.0μl DNA amplified) with a 6μl Identifiler fast PCR amplification in which the combining annealing/extension temperature was increased to 63°C and the final extension was increased to 20min, with subsequent detection with POP-4/36cm array using a 3kV 10sec. These conditions eliminated NSA and -A.
162
When a 63°C annealing/extension temperature was tested using 20 samples, not only was all
non-specific amplification eliminated (see Figure 3.36), but n-8 stutter was eliminated as well,
without compromising profile quality: full profiles were obtained from 95% of samples (same as
that obtained using 61°C); average peak height was 1093rfu (decreased from 1755rfu using
61°C, but still within the target ranged); average PHR was 82.3% (84.6% using 61°C); 0.8% of loci
and 5% of samples had PHR <50% (same as 61°C); and average CV of LPH:TPH was 0.327 (0.336
using 61°C).
3.3 Development of a Normalized Extraction using the ChargeSwitch® Forensic DNA Purification Kit
3.3.1 Initial Evaluation
Actual DNA concentrations fell within the predicted maximum range for 56% of samples,
based upon the advertised maximum bead binding capacities of 5-10mg genomic DNA per
nanogram beads (see Table 3.10). Actual bead binding capacities ranged from 2.4-9.1mg/ng,
averaging about 5.5mg/ng. Of the nine samples, only one (0.5µl beads/80µl elution) fell within
the target amplification range for fast and standard PCR, while two additional (0.5µl beads/60µl
elution and 1µl beads/80µl elution) may have been suitable for fast PCR but not standard, and
one other (0.25µl beads/80µl elution) may have been suitable for standard but not fast PCR.
These four samples, as well as one additional obtained from using 0.25µl beads and a 60µl
elution volume with a slightly low DNA concentration (0.25ng/µl), were all amplified using 3µl
reactions for HS, ID+ and ID fast (using 0.9µl DNA) and standard (using 1.2µl DNA) PCR. For all
samples, full profiles were obtained. As expected, peak heights were higher for samples with
greater DNA input and PHR <50% were problematic for samples with low (0.25µl beads/60µl
163
elution/HS, ID+, ID fast and 0.25µl beads/80µl elution/ID standard) or high (1µl beads/80µl
elution/ID fast) amounts of DNA. Pull up was also more prevalent in samples with greater DNA
input as well.
Table 3.10
Summary of DNA Recoveries
60µl Elution Volume 80µl Elution Volume
Volume (µl) of
Purification Buffer:Beads
Predicted Maximum
[DNA] (ng/µl)
Actual [DNA] (ng/µl)
Bead Binding Capacity (mg/ng)
DNA Yield (ng)
Sample Source
Predicted Maximum
[DNA] (ng/µl)
Actual [DNA] (ng/µl)
Bead Binding Capacity (mg/ng)
DNA Yield (ng)
Sample Source
100:5.0 10.4 – 20.8 13.4 6.75 804 Top of
swab (¼) 7.8 – 15.6 3.87 2.60 310
Side of swab (¼)
100:2.0 4.2 – 8.3 N/A N/A N/A N/A 3.1 – 6.3 4.83 7.88 386 Bottom rim of swab
100:1.0 2.1 – 4.2 3.74 9.06 224 Top of
swab (¼) 1.6 – 3.1 1.26a 4.06 101
Side of swab (¼)
100:0.50 1.0 – 2.1 1.30a 6.29 78.2 Top of
swab (¼) 0.8 – 1.6 0.914b 5.88 73.1
Side of swab (¼)
100:0.25 0.50 – 1.0 0.25 2.41 15.0 Top of
swab (¼) 0.4 – 0.8 0.384c 4.93 30.7
Side of swab (¼)
Note. [DNA] = DNA concentration. Actual DNA concentrations were either within the predicted maximum range or below. Since fast PCR utilizes 0.9µl DNA and standard PCR utilizes 1.2µl DNA, DNA concentration should be 0.417-1.25ng/µl for a sample to be amplified within the 0.375-1.50ng target for fast and standard PCR. aWithin target for fast PCR, but not standard.
bWithin target for fast and standard PCR.
cWithin target for standard
PCR, but not fast.
3.3.2 Optimization of Normalized Extraction for Buccal Swab Cuttings and Buccal DNA Collector™ Punches
Using a small sample set of swabs and punches from four individuals (one high shedder,
two moderate shedders and one low shedder) processed in triplicate for each of five bead
volumes (0.25µl, 0.50µl, 1.0µl, 2.0µl and 5.0µl) and various elution volumes (60µl and 80µl for
swabs and 120µl for punches), DNA concentrations expectedly increased as more beads were
used. In general DNA concentrations were more often than not lower than the predicted
maximums for punches and swabs, but reached predicted maximums more often with swabs
164
(48% of samples) than punches (18% of samples). Additionally, more variation was seen
between individuals than within an individual.
For punches, the maximum predicted concentration was rarely obtained. Maximum
concentrations were never obtained with the use of 5.0µl beads, the third punch (except from
the high shedder with 0.25µl beads), the low shedder or one of the moderate shedders.
Maximum predicted concentrations were obtained for replicate punches 1 and 2 with 0.25-
2.0µl beads 100% of the time from the high shedder and 50% of the time from the other
moderate shedder. No single extraction method resulted in all samples’ DNA concentrations
falling within the targeted range (0.208-0.833ng/µl) for fast and standard PCR to achieve DNA
inputs within the optimal range (0.375-1.50ng). Therefore, a total of ten samples (the highest
and lowest concentration from each individual using 0.5µl beads, as well as the overall highest
and lowest concentrated samples using 1.0µl beads) were chosen for 6µl Identifiler fast and
standard amplification. These samples (and 9947A positive amplification control) unexpectedly
exhibited severe -A; the amplification was repeated and -A was confirmed. Furthermore, full
profiles were only obtained from the high shedder, while all of the other samples exhibited
signs of PCR inhibition due to too much DNA or some other kind of inhibitor in the extract. To
compensate for this apparent inhibition, less DNA (1.0µl) was used in fast and standard PCR
amplifications for 0.5µl and 1.0µl bead samples, and injection time was increased from 7sec to
10sec (at 3kV); this eliminated -A, but overall peak heights were too low and profiles suffered
from dropout.
For swabs, the maximum predicted concentration was obtained about half (48%) the
time, more so from the 80µl elution volume (58%) than the 60µl (38%). Maximum
165
concentrations were obtained less often with the use of 1.0µl (40%), 2.0µl (46%) or 5.0µl (10%)
beads compared to 0.25µl (79%) and 0.50µl (75%) beads. As seen with punches, the third swab
cutting nearly always (95%) had a lower DNA concentration than the first two cuttings, but its
concentration wasn’t always less than the predicted maximum. Predicted maximums were
obtained by the high shedder most often (80%), followed by one of the moderate shedders
(57%), the low shedder (30%) and lastly, the other moderate shedder (20%). As with the
punches, no single extraction method resulted in all samples’ concentrations falling with the
targeted DNA concentration (0.417-1.25ng/µl) for fast and standard PCR to achieve DNA inputs
within the optimal range (0.375-1.50ng). The percentage of samples within range was highest
from the 0.25µl bead/60µl elution data set (83%; average concentration of 0.921ng/µl),
followed by the 0.50µl bead/80µl elution data set (75%; average concentration of 0.535ng/µl).
All 12 samples from these two data sets were amplified with 3µl fast (0.9µl DNA) and standard
(1.2µl DNA) HS, ID+ and ID amplifications. Full profiles were obtained from 100% of 0.25µl and
0.50µl bead samples using fast and standard amplifications, but peak heights were lower than
desired for many 0.25µl bead profiles and PHR <50% were present in nearly every amplification
data set (8% of samples per data set). Oversaturated peaks were problematic for standard
amplification of the 0.50µl bead data set, likely due to the increased DNA inputs compared to
the 0.25µl data set, as well as fast amplifications (standard amplifications have 33% more DNA
than fast amplifications).
Overall, this small sample size for punches and swabs was very informative. Though the
extraction procedures varied only with regard to bead and elution volumes, PCR inhibition
appeared to be a problem for punch samples, while a balancing act was needed for peak
166
heights for swab samples. As studies progressed, it became difficult to determine if poor punch
profiles were due to poor sample collection, the normalized extraction, the 6µl Identifiler fast
PCR amplification and/or a combination of these variables. In an effort to reduce the presence
of a possible inhibitor being carried over into the extract, different amounts of wash buffer
(125µl for both washes; 250µl for the first wash and 125µl for the second wash; and 250µl for
both washes) were compared, but did not improve extract quality.
Inconsistent -A with the 6µl Identifiler fast protocol and low-level, non-specific
amplification peaks for 3µl and 6µl Identifiler fast protocols resulted in post-validation
modifications of both fast protocols. The 3µl and 6µl Identifiler fast protocols were reassessed
for annealing temperature (changed to 63°C) to eliminate low-level NSA that had begun to
appear, and the 6µl protocol was also reassessed for final extension time (changed to 20min) to
reduce -A formation (see section 2.3.5 Post-Validation Modifications to Identifiler Fast PCR).
Additionally, comparisons between fast and standard amplification became difficult due to the
different volumes of DNA allotted for each amplification, therefore later side-by-side
comparisons of fast and standard amplifications utilized the same volume of DNA for a
particular experiment. These amplification modifications were helpful for the extraction
normalization development process by increasing the upper end of the optimal DNA input
range to >1.50ng, but obviously did not solve all the problems (e.g., PCR inhibition).
Eventually, a large data set of swab samples unexpectedly exhibited poor results that
also suggested inhibition due to the normalized extraction. At that point, the extraction
instrument was identified as being out of calibration, which had been resulting in bead
carryover into the elution plate for an unknown amount of time. A different extraction
167
instrument was used from that point on and no bead carryover was ever noted; swab profile
quality improved and another large data set utilizing 0.5µl beads and 60µl elution volume
confirmed that the procedure was ready for validation (first pass success rates were >90% for
all three 3µl fast PCR methods and both detection methods).
Punch profile quality improved after bead carryover into the elution plate was
eliminated, but was still not ideal when a large data set was tested using 0.5µl beads and 60µl
elution volume (first pass success rates were 77-86%). Next, several small (n=10), focused
studies ensued regarding the lysis step in the normalized extraction: incubation time was
increased from 1.5hr to overnight; Proteinase K volume was increased from 5.0µl to 10µl; and a
post-incubation vortex step was assessed. Overnight and 1.5hr incubation samples were
amplified using 0.50µl, 1.0µl and 1.8µl DNA; for all data sets, overnight incubation resulted in
dramatically decreased DNA concentrations (0.34ng/µl), as well as a reduction in full profiles
(50-80%3) and pass rates (40-70%2), compared to 1.5hr incubation (0.51ng/µl, 60-90% and 50-
90%, respectively). Use of 10µl Proteinase K demonstrated no significant differences compared
to the use of 5.0µl with regard to DNA concentrations (average of 1.03ng/µl and 1.04ng/µl,
respectively), average percent of alleles detected (96% and 97%, respectively), full profiles (80%
and 90%, respectively) or pass rates (80% for both). Lastly, the post-incubation vortex step
unexpectedly resulted in contamination and was not pursued further.
Given that these evaluations did not result in changes to the extraction parameters,
another small data set was processed using 0.50µl beads/60µl elution volume and 1.0µl
beads/80µl elution volume for a comparison to the data set that was previously tested. As
3 The 0.5µl data set was excluded from percentage of full profiles and pass rates because no full profiles were
obtained using either an overnight or 1.5hr incubation.
168
expected, DNA concentrations were higher from the 1.0µl data set, with an average of
1.75ng/µl compared to 1.15ng/µl for the 0.50µl data set. Samples were amplified using 1.0µl
and 1.8µl DNA in 6µl Identifiler fast amplifications and resulted in pass rates of 70-100% after
injection parameters were reduced from 3kV 10sec to 2kV 7sec to compensate for higher peak
heights. A large data set of punch samples was extracted using 1.0µl beads and 80µl elution
volume, followed by amplification using 6µl Identifiler fast PCR, as well as 3µl fast PCR reactions
(HS/ID+/ID). Though pass rates (85-91%) were not as high as desired for all methods (>90%),
time and monetary constraints forced the project to move into the validation phase.
3.3.3 Validation of Normalized ChargeSwitch Extractions Within the New First Pass Processes
3.3.3.1 Sensitivity
DNA concentrations from normalized extraction of 170 buccal swab samples and 70
punch samples were 0.130-2.52ng/μl and 0.020-3.67ng/µl, respectively. Using 0.9μl DNA for 3μl
reaction volumes (swabs and punches) and 1μl DNA for 6μl reactions (punches only), 0.117-
2.268ng, 0.018-3.30ng and 0.020-3.67ng were amplified using 3μl fast PCR methods (all three
kits) for swabs and punches and 6μl Identifiler fast for punches only, respectively. All samples
were processed with POP-4 and POP-6 detection, except that 84 of the 3μl fast PCR swab
samples (all three fast PCR methods) were only detected on POP-6, thereby decreasing the
tested range for POP-4 to 86 samples (0.477-2.268ng) for 3µl amplifications of swabs.
169
A) Swabs (3µl Fast Amplifications)
B) Punches (3µl/6µl Fast Amplifications)
Figure 3.37. Tested sensitivity ranges for normalized extraction coupled with fast PCR. Samples exhibiting allelic dropout are displayed for swab (n=84 for POP-6 test range and 86 for POP-4 test range)(A) and punch (n=70)(B) samples.
Full profiles were obtained over the vast majority of DNA inputs for swab samples for all fast
amplifications (regardless of detection method), with dropout only occurring when 0.135ng
was amplified using PowerPlex 16 HS or Identifiler Plus or when 0.117ng was amplified using
Identifiler (see Figure 3.37). Full profiles were also obtained over the vast majority of DNA
inputs for punch samples, but more dropout was observed with punches than with swabs.
Furthermore, it is interesting that full profiles were obtained from less than 50% of punch
0.00 0.50 1.00 1.50 2.00 2.50
DNA Input (ng)
HSF POP-4
HSF POP-6
ID+F POP-4
ID+F POP-6
IDF POP-4
IDF POP-6
3µl Test Range POP-4
3µl Test Range POP-6
0.00 1.00 2.00 3.00 4.00
DNA Input (ng)
HSF POP-4HSF POP-6ID+F POP-4ID+F POP-6IDF POP-4IDF POP-63µl Test Range POP-4 / POP-6IDF6 POP-4IDF6 POP-66µl Test Range POP-4 / POP-6
170
samples within the 0.75-1.00ng range for all 3µl fast amplifications, despite the fact that nearly
100% full profiles were obtained from DNA input ranges greater than and less than 0.75-1.00ng
(see Figure 3.38). Thus, dropout was isolated to low template swab samples, but was more
sporadic for punch samples; upper limits to sensitivity could not be established and remain to
be determined. Sensitivity ranges based upon these data are summarized in Table 3.11 and are
very similar to those determined from the fast PCR validations.
A) Swabs (3µl Fast Amplifications)
B) Punches (3µl/6µl Fast Amplifications)
Figure 3.38. Percent full profiles obtained using normalized extraction coupled with fast PCR. Data from swabs and punches are displayed across the tested sensitivity ranges. It should be noted that sample size varies for each DNA input range. 1n=0 (POP-4) and n=2 (POP-6). 2n=0 (POP-4) and n=7 (POP-6). 3n=15 (POP-4) and n=33 (POP-6). 4n=65 (POP-4) and n=111 (POP-6). 5n=6 (POP-4) and n=17 (POP-6). 6n=2 (3µl and 6µl). 7n=1 (3µl and 6µl). 8n=4 (3µl) and n=1 (6µl). 9n=22 (3µl) and n=19 (6µl). 10n=37 (3µl) and n=42 (6µl). 11n=2 (3µl) and n=3 (6µl).
0%
20%
40%
60%
80%
100%
0.00-0.188ng 0.188-0.375ng 0.375-0.750ng 0.750-1.50ng 1.50-3.00ng
HSF POP-4
HSF POP-6
ID+F POP-4
ID+F POP-6
IDF POP-4
IDF POP-61 2 3 4 5
0%
20%
40%
60%
80%
100%
0.00-0.188ng 0.188-0.375ng 0.375-0.750ng 0.750-1.50ng 1.50-3.00ng 3.00-3.75ng
HSF POP-4
HSF POP-6
ID+F POP-4
ID+F POP-6
IDF POP-4
IDF POP-6
IDF6 POP-4
IDF6 POP-66 7 8 9 10 11
171
Table 3.11
Sensitivity Range for Normalized Extraction and Fast PCR
Sample Type Fast PCR Process Sensitivity Range
Swabs PowerPlex 16 HS (3µl)
Identifiler Plus (3µl) Identifiler (3µl)
0.188-3.00ng 0.188-3.00ng 0.188-3.00ng
Punches
PowerPlex 16 HS (3µl) Identifiler Plus (3µl) Identifiler (3µl) Identifiler (6µl)
0.188-3.75ng 0.188-3.75ng 0.188-3.75ng 0.188-3.75ng
Note. Upper limits to the sensitivity ranges may be even higher for all processes. Fast PCR validations using samples extracted with the current, non-normalized extraction procedure identified sensitivity ranges to be 0.375-3.00ng for 3µl Identifiler and 0.188-3.00ng for the other three fast PCR procedures listed above.
3.3.3.2 Reproducibility
DNA concentrations obtained from duplicate buccal cell samples were not significantly
different when processed within (swabs or punches) or between (swabs) batches (two-tailed,
match paired t-test; p=0.16, p=0.29 and p=0.31, respectively); see Table 3.12 for a summary of
DNA concentrations. Furthermore, absolute values of differences observed between duplicates
averaged 0.27ng/μl, 0.45g/µl and 0.23ng/μl for swabs within a batch, punches within a batch
and swabs between batches, respectively (see Table 3.13). Though differences between
duplicates were signficantly greater between batches than within (Tukey HSD, p=5.7x10-4 and
p=4.1x10-3 for swabs between batches versus punches within batch and swabs between
batches versus swabs within batch, respectively), the degree of reproducibility was acceptable
from all duplicates.
172
Table 3.12
Reproducibility of DNA Concentrations Within and Between Normalized Extraction Batches
Within a Batch Between Batches
[DNA] (ng/µl)
Swabs Punches Swabs
Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2
Average 1.14 1.21 1.71 1.76 1.03 1.12 95% CI [1.03, 1.25] [1.11, 1.32] [1.48, 1.94] [1.52, 2.01] [0.928, 1.13] [0.943, 1.29]
Minimum 0.630 0.530 0.020 0.030 0.370 0.130 Maximum 2.52 2.27 3.30 3.67 1.95 2.05
Note. [DNA] = DNA concentration. Reproducibility of punches between batches was not tested.
Table 3.13
Differences Between Duplicates Within and Between Normalized Extraction Batches
Within a Batch Between Batches
∆[DNA] (ng/µl)
Swabs Punches Swabs
Average 0.27a 0.23a 0.45 95% CI [0.21, 0.33] [0.16, 0.29] [0.35, 0.54]
Minimum 0.01 0.01 0.02 Maximum 0.81 0.77 1.23
Note. ∆[DNA] = difference in DNA concentration between replicate samples. Reproducibility of punches between batches was not tested. Averages sharing a common subscript are not statistically different at α=0.05 according to the Tukey HSD procedure.
3.3.3.3 Comparison to Current Process
Maximum bead binding capacities are advertised as 5-10mg genomic DNA per
nanogram beads using the ChargeSwitch extraction procedure as described by the
manufacturer. As expected, actual bead binding capacities (see Table 3.14) significantly
increased via the use of the normalized extraction procedure compared to that of the current,
non-normalized procedure for high quality swabs, old swabs and punches (two-tailed, match-
pair t-test; p=1.24x10-24, 5.30x10-4 and 6.00x10-15, respectively). Average capacities nearly
173
doubled using normalized extraction for all sample types, but often did not reach the
advertised maximum binding capacity.
Representative profiles from the two new first pass options (including normalized
extraction, fast PCR and POP-4/POP-6 detection) and the current first pass procedures are
displayed in Figures D.11 – D.14.
Table 3.14
Bead Binding Capacities
Bead Binding Capacity
High Quality Swabs Old, Degraded Swabs Punches
Normalized Non-
Normalized Normalized
Non- Normalized
Normalized Non-
Normalized
Average 5.65 2.50 4.44 1.93 4.17 2.12 95% CI [5.27, 6.03] [2.14, 2.87] [3.82, 5.06] [1.25, 2.60] [3.77, 4.57] [1.85, 2.39]
Minimum 2.54 0.371 1.87 0.384 0.050 0.032 Maximum 12.1 8.48 7.50 6.36 8.80 6.52
% Reaching Maximum
62% 10% 32% 9.0% 31% 1.4%
Note. Bead binding capacities are displayed as milligram DNA per nanogram beads.
3.3.3.3.1 High Quality Swabs
All of the high quality buccal swab samples processed under the various conditions
yielded STR profiles consistent with those obtained with the current procedure, and no
unexplained alleles were observed. Of the 86 samples processed, 100% yielded full profiles for
each of the data sets, except from the data set utilizing normalized extraction, 3µl Identifiler
standard PCR and POP-4 detection, from which full profiles were obtained from 94% of the
samples, with an average of 99% alleles detected per sample (see Table 3.15). The Identifiler
samples that experienced dropout via standard amplification had DNA inputs of 1.1-2.3ng and
exhibited signs of PCR inhibition, presumably due to too much DNA.
174
Table 3.15
Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: High Quality Swabs
PowerPlex 16 HS (3µl) Identifiler Plus (3µl) Identifiler (3µl)
Normalized Current Process
Normalized Current Process
Normalized Current Process
Fast Standard Fast Standard Fast Standard
Profile Assessment POP-4 POP-6 POP-4 POP-4 POP-6 POP-4 POP-4 POP-6 POP-4
Full Profiles Detected 100%a 100%a 100%a 100%a 100%b 100%b 100%b 100%b 100%c 100%c 94% 100%c Alleles Detected
Average 100%d 100%d
100%d 100%d 100%e 100%e 100%e 100%e 100%f 100%f
98.9% 100%f 95% CI [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [100%, 100%] [97.9%, 100%] [100%, 100%]
Peak Height (rfu) Average 1495g 1339g 1762 1246 1226h 1218h 1251h 2062 1219i 1176 1261i 2030 95% CI [1462, 1528] [1309, 1368] [1722, 1801] [1218, 1273] [1201, 1251] [1192, 1244] [1226, 1276] [2028, 2097] [1193, 1245] [1148, 1204] [1230, 1292] [1985, 2075] Observed Maximum 5946 5503 7903 4750 4437 4592 4644 5572 4645 4761 4596 6518 Observed Minimum 349 351 326 190 381 355 437 706 365 219 178 476
Inter-Locus Balance: CV of LPH:TPH
Average 95% CI
0.360 [0.351, 0.368]
0.324 [0.317, 0.331]
0.402 [0.390, 0.413]
0.442 [0.430, 0.455]
0.290j [0.281, 0.298]
0.331 [0.322, 0.339]
0.285j [0.272, 0.298]
0.255 [0.248, 0.262]
0.315 [0.301, 0.330]
0.353 [0.338, 0.368]
0.437k [0.411, 0.463]
0.418k [0.407, 0.429]
Intra-Locus Balance: PHR
Average 95% CI PHR <50%p
0.870l,m [0.864, 0.875]
0%q
0.864 l [0.858, 0.870]
0%q
0.888m [0.883, 0.893]
0.096%q
0.885l,m [0.880, 0.890]
0.096%q
0.876n [0.871, 0.881]
0%r
0.872n [0.867, 0.878]
0%r
0.889n [0.885, 0.894]
0%r
0.890n [0.885, 0.895]
0.097%r
0.877o [0.872, 0.882]
0%s
0.873o [0.868, 0.879]
0%s
0.875 o [0.869, 0880]
0.097%s
0.87 o [0.867, 0.878]
0.097%s First Pass Success Rate
100% t 99% t 99% t 99%t 97%u 99%u 64% 97%u 99%v 95%v 92%v 98%v
Failure Reasons Dropout PHR <50%w Stutter (n-4) -A Failed to Transferx Injection Failure Loss of Resolution
0% 0% 0% 0% 0% 0% 0%
0% 0% 0% 0% 0%
1.2% 0%
0%
1.2% 0% 0% 0% 0% 0%
0%
1.2% 0% 0% 0% 0% 0%
0% 0%
1.2% 0% 0%
1.2% 1.2%
0% 0% 0% 0% 0%
1.2% 0%
0% 0% 0%
36% 0% 0% 0%
0%
1.2% 0%
1.2% 0%
1.2% 0%
0% 0%
1.2% 0% 0% 0% 0%
0% 0% 0% 0% 0%
4.7% 0%
5.8% 1.2% 0%
1.2% 0% 0% 0%
0%
1.2% 0% 0%
1.2% 0% 0%
Oversaturationy 0% 0% 3.5% 0% 0% 0% 0% 0% 0% 0% 0% 1.2%
Note. Profile summaries are displayed for buccal swab samples using new versus current first pass options. Averages (within a primer set) sharing a common subscript are not statistically different at α=0.05 according to the Tukey HSD procedure. Frequencies of full profiles detected, first pass success and PHR <50% at all heterzygous loci (within a primer set) sharing a common subscript are not statistically different at α=0.05 according to the Chi-square contingency test. pPercent of all heterozygous loci.
wPercent of samples.
xTemplate DNA failed to transfer from extraction plate to amplification plate.
yNot enough
oversaturation to result in failure.
175
Overall, average peak height from high quality swabs ranged from 1176rfu to 2062rfu
and was somewhat higher using the normalized extraction and the three downstream process
combinations for PowerPlex 16 HS profiles compared to those obtained using the current first
pass process, but the opposite was observed for Identifiler Plus and Identifiler. Observed
maximum peak heights ranged from 4437rfu to 7903rfu and only exceeded 6000rfu (a typical
starting point for laser oversaturation) for two of the 12 data sets. Observed minimum peak
heights ranged from 178rfu to 706rfu.
Average heterozygote peak height ratio ranged from 86.4% to 89.0% for all data sets,
but were slightly higher (1-2%) for PowerPlex 16 HS and Identifiler Plus standard PCR data sets
(normalized and current extraction) compared to fast PCR. Interestingly, the only profiles to
exhibit PHR <50% were from standard amplifications (normalized and current extraction), but
were minimal, nonetheless, representing 1.2% of samples from their respective data sets.
In addition to assessing intra-locus peak balance via peak height ratios, inter-locus peak
balance was assessed through the coefficient of variation (CV) of each sample’s LPH:TPH ratios.
Average CV values ranged from 0.255 to 0.442. In general, PowerPlex 16 HS and Identifiler
profiles obtained via standard amplification (normalized and current extraction) showed more
inter-locus imbalance than fast PCR. The opposite was true for Identifiler Plus, in which profiles
obtained via fast PCR were just slightly more imbalanced than standard amplification, more so
from normalized extraction than the current extraction procedure. However, Identifiler Plus
inter-locus balance using fast PCR was as good as or better than that observed for PowerPlex 16
HS and Identifiler fast PCR, all of which were <0.360.
176
One additional item of note is that during the validation of Identifiler fast PCR, a low
level artifact at Amelogenin was noted in 0.7% of the samples. Despite an increase in annealing
temperature, this artifact was again observed in one sample (1.2%) with a height of 76rfu using
POP-6 detection; it was also present using POP-4 detection of the same amplification product,
but was below threshold.
Samples were given a pass/fail rating based on the first pass analysis guidelines, and
pass rates were 92-100% for all except Identifiler Plus standard amplification of normalized
samples. Several noteworthy trends were observed. First, the two new first pass options
involving normalized extraction and fast PCR yielded as high or higher pass rates than the
current first pass procedure from the corresponding amplification kit. Second, the two new first
pass options also yielded as high or higher pass rates than the intermediate processes involving
normalized extraction followed by standard amplification and traditional POP-4 detection,
suggesting inferior performance of standard amplification coupled with the normalization
extraction procedure, especially with the high rate of detected -A peaks observed with
Identifiler Plus, as well as the relatively high rate of dropout observed with Identifiler. Third,
POP-6 detection appeared to be more prone to injection failures than POP-4, but it should be
noted that the 3130xl instrument setup with POP-6 was not used as frequently as the POP-4
3130xl instrument, and it has been noted during production that infrequent/non-routine use of
a 3130xl can result in increased injection issues (personal observation). Lastly, additional
artifacts (including n-8 stutter, elevated baseline and spikes) and/or minor profile defects
(oversaturation) that were present but within the guidelines for a passing profile are presented
in Figure 3.39 or Table 3.15. These indicate an increase in detectable n-8 stutter peaks for
177
PowerPlex 16 HS and elevated baseline peaks for all three amplification kits compared to the
current process.
Positive and negative amplification controls performed as expected. It should be noted
that of the 1032 STR profiles obtained from this particular study, nine samples experienced
detection issues using either POP-4 or POP-6. Those experiencing loss of resolution or injection
failure were excluded from calculations regarding profile completeness, peak height and PHR,
but were included in pass/fail rates. Those experiencing incomplete data collection (i.e., the
450bp peak of the LIZ internal size standard was not collected during detection due to room
temperature fluctuations) were analyzed without the 450bp peak of the size standard.
Figure 3.39. Detection of non-failing artifacts using current and normalized extractions. These did not result in failing profiles (see first pass analysis guidelines).
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
POP-4 POP-6 POP-4 POP-4 POP-6 POP-4 POP-4 POP-6 POP-4
Fast PCR Std PCR Current Fast PCR Std PCR Current Fast PCR Std PCR Current
Normalized Extraction Process Normalized Extraction Process Normalized Extraction Process
HS ID+ ID
Stutter (n-8) Elevated Baseline Spikes
178
3.3.3.3.2 Old, Degraded Swabs
All of the old degraded buccal swab samples processed using Identifiler Plus fast PCR
with POP-6 detection yielded STR alleles/profiles consistent with those obtained with the
current procedure. Of the 22 samples processed, 95% yielded full profiles with the normalized,
fast procedure, compared to 91% with current first pass process, with an average of 97.4% and
98.2% of alleles detected, respectively (see Table 3.16). Dropout tended to occur due to alleles
of larger loci falling below threshold.
Overall, average peak height was lower using the normalized, fast procedure compared
(794rfu) to that obtained using the current first pass process (1647rfu), which is somewhat
counter-intuitive given that dropout rates were higher from the current first pass procedure.
Observed maximum peak heights were 4172rfu and 6855rfu, while observed minimum peak
heights were 76rfu and 80rfu, respectively. These are all lower than peak heights obtained from
swabs up to four months old.
Average heterozygote peak height ratio was 85.2% and 84.7% for the normalized, fast
procedure and current first pass procedure, with 9.1% and 4.6% of profiles exhibiting PHR
<50%, respectively. Average CV of LPH:TPH values were 0.490 and 0.404, respectively. Intra-
locus and inter-locus balance decreased for these old, degraded swabs compared to swabs that
were up to four months old using both first pass procedures. Furthermore, both processes had
86% pass rates, with profile failures limited to dropout and PHR <50%. No additional
noteworthy artifacts were observed. Positive and negative controls performed as expected.
179
Table 3.16
Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: Old,
Degraded Quality Swabs
Identifiler Plus (3µl)
Profile Assessment Normalized Extraction, Fast
PCR, POP-6 Current Process
Full Profiles Detected 95% 91% Alleles Detected
Average
97.4% 98.2% 95% CI [92.3%, 100%] [95.3%, 100%]
Peak Height (rfu) Average 794* 1647* 95% CI [748, 840] [1556, 1738] Observed Maximum 4172 6855 Observed Minimum 76 80
Inter-Locus Balance: CV of LPH:TPH Average 95% CI
0.490**
[0.465, 0.515]
0.404**
[0.376, 0.432] Intra-Locus Balance: PHR
Average 95% CI PHR <50%a
0.852
[0.838, 0.865] 0.76%
0.848
[0.835, 0.861] 1.0%
First Pass Success Rate 86% 86% Failure Reasons
Dropout PHR <50%b
4.6% 9.1%
9.1% 4.6%
Oversaturationc 0% 0% Note. Profile summaries are displayed for old, degraded swab samples using new versus current first pass procedures. aPercent of all heterozygous loci.
bPercent of samples.
cNot enough oversaturation to result in failure.
*p=8.11x10-85
; two-tailed, match paired t-test. **p=1.16x10-7
; two-tailed, match paired t-test.
3.3.3.3.3 Buccal DNA Collector™ Punches
All of the buccal punch samples processed using Identifiler Plus fast PCR with POP-6
detection yielded STR alleles/profiles consistent with those obtained with the current
procedure. Of the 68 samples processed (two of the original 70 samples were removed due to
processing error), full profiles were obtained from 84-96% of samples, with an average of 93.1-
180
97.9% of alleles detected (see Table 3.17). As was seen with the old, degraded samples,
dropout often occurred at larger sized loci, but some profiles also exhibited signs of PCR
inhibition (more so from the 6µl Identifiler fast PCR amplifications than the others).
Average peak height from punches was 821-2430rfu and was higher from each of the
3µl current first pass processes than their normalized alternatives, but the opposite was seen
for 6µl Identifiler first pass options. Observed maximum peak heights ranged from 2941rfu to
8055rfu, while observed minimum peak heights ranged from 75rfu to 209rfu. These minimum
peak heights were more similar to those obtained from old, degraded buccal swabs than those
from high quality swabs. Furthermore, a wider range of peak heights was observed for several
of the first pass procedures for punches (e.g., 106-8055rfu, 75-7575rfu, etc.) compared to high
quality swabs, making it difficult to rely on injection parameter adjustments to fine tune peak
height. Additionally, oversaturated peaks were more abundant in these data sets (more so from
the current first pass procedures), but never resulted in profile failure.
Average heterozygote peak height ratio was 85.2-87.3% for all punch data sets and
generally did not differ significantly for first pass methods using the same primer set and
reaction volume. As seen with high quality swabs, more samples exhibited PHR <50% with the
current first pass process (1.5-7.4%) than with normalized, fast PCR processes (0-1.5%). Average
CV of LPH:TPH values were 0.268-0.502 and tended to exceed the desire level ( 0.350) for all
3µl Identifiler first pass options and both normalized, 6µl Identifiler fast PCR first pass options.
Intra-locus and inter-locus balance were roughly comparable between high quality swabs and
punches for the respective first pass procedures, but inter-locus balance was considerably
better from punch samples (0.354) than swabs (0.442) using the current first pass options.
181
Table 3.17
Comparison Between Normalized and Non-Normalized Extraction First Pass Processes: Punches
PowerPlex 16 HS (3µl) Identifiler Plus (3µl) Identifiler (3µl) Identifiler (6µl)
Normalized, Fast Current Process
Normalized, Fast Current Process
Normalized, Fast Current Process
Normalized, Fast Current Process Profile Assessment POP-4 POP-6 POP-4 POP-6 POP-4 POP-6 POP-4 POP-6
Full Profiles Detected 91%a 96%a 90%a 91%b 91%b 91%b 91%c 88%c 84%c 90%d 87%d 88%d Alleles Detected
Average
96.4%e 95.7%e 97.1%e 95.6%f 95.3%f 97.9%f 96.4%g 93.1%g 93.5%g 95.8%h 95.9%h 95.7h 95% CI [93.4%, 99.4%] [92.1%, 99.2%] [94.3%, 100%] [91.7%, 99.5%] [91.4%, 99.2%] [95.8%, 100%] [93.3%, 99.5%] [88.1%, 98.1%] [88.9%, 98.1%] [92.3%, 99.2%] [92.9%, 99.0%] [92.2%, 99.2%]
Peak Height (rfu) Average 1527 1275 1930 1619 1294 2430 1573i 1412 1619i 1140 1244 821 95% CI [1491, 1563] [1243, 1308] [1879, 1981] [1577, 1661] [1257, 1330] [2378, 2482] [1530, 1615] [1360, 1463] [1573, 1664] [1099, 1181] [1200, 1288] [801, 841] Observed Maximum 4999 4784 8055 7138 5754 7438 7303 6715 6533 7575 6167 2941 Observed Minimum 169 113 106 123 86 209 127 164 98 75 78 81
Inter-Locus Balance: CV of LPH:TPH
Average 95% CI
0.342j [0.330, 0.354]
0.334j [0.320, 0.348]
0.354j [0.336, 0.372]
0.340 [0.322, 0.358]
0.383 [0.364, 0.401]
0.268 [0.249, 0.286]
0.374k [0.354, 0.393]
0.410k,l [0.390, 0.431]
0.438l [0.409, 0.466]
0.481m [0.452, 0.510]
0.502m [0.479, 0.524]
0.276 [0.259, 0.293]
Intra-Locus Balance: PHR
Average 95% CI PHR <50%r
0.868n [0.862, 0.874]
0.52%s
0.864n [0.858, 0.870]
0.52%s
0.861n [0.854, 0.869]
1.0%s
0.873o [0.867, 0.880]
0%t
0.868o [0.862, 0.874]
0.13%t
0.876o [0.870, 0.882]
0.50%t
0.871p [0.865, 0.876]
0%u
0.864p [0.858, 0.870]
0%u
0.853 [0.844, 0.861]
1.6%
0.862q [0.855, 0.868]
0.26%v
0.856q [0.850, 0.862]
0.39%v
0.852q [0.845, 0.859]
0.90%v First Pass Success Rate 90%w 90%w 85%w 91%x 91%x 90%x 91%y 88%y,z 76%z 90%aa 85%aa 82%aa Failure Reasons
Dropout PHR <50%ab Stutter (n-4)
8.8% 1.5% 0%
8.8% 1.5% 0%
10% 4.4% 0%
8.8% 0% 0%
8.8% 0% 0%
8.8% 1.5% 0%
8.8% 0% 0%
12% 0% 0%
16% 7.4% 0%
10% 0% 0%
13% 0%
1.5%
12% 5.9% 0%
Oversaturationac 0% 0% 12% 0% 0% 12% 0% 0% 1.5% 1.5% 0% 0%
Note. Profile summaries are displayed for buccal swab samples using new versus current first pass options. Averages (within a primer set) sharing a common subscript are not statistically different at α=0.05 according to the Tukey HSD procedure. Frequencies of full profiles detected, first pass success and PHR <50% at all heterzygous loci (within a primer set) sharing a common subscript are not statistically different at α=0.05 according to the Chi-square contingency test. rPercent of all heterozygous loci.
abPercent of samples.
acNot enough oversaturation to result in failure.
182
First pass success rates for punches were lower than those observed for high quality
swabs, ranging from 76% to 91%. Dropout was the primary cause for failure for all first pass
options and accounted for 8.8-16% of samples, which was substantially higher than the 0-5.8%
observed with high quality swabs. Furthermore, dropout more often occurred using the current
first pass procedures than the new first pass alternatives. PHR <50% was the next leading cause
for profile failures and again accounted for a larger percentage of punches (0-7.4%) than swabs
(0-1.2%), as well as a higher occurrence among current than new first pass procedures
(discussed above). Lastly, a single sample (1.5% of the normalized, 6µl Identifiler fast PCR,
POP-6 data set) failed due to high n-4 stutter.
3.3.3.4 Contamination Assessment
No sample or reagent blank exhibited any signs of contamination for either normalized
extraction procedure during the contamination assessment. It so be noted (as previously
discussed during the development section) that applying a brief 5sec vortex step following the
56°C incubation repeatedly resulted in low-level, well-to-well contamination and was removed
from the extraction procedure.
183
CHAPTER 4
CONCLUSIONS
4.1 Improvements to Processing Efficiency
The goal of this research project was to improve the efficiency of an existing database
workflow within an established DNA crime laboratory via development and implementation of
a normalized extraction procedure to eliminate the need for quantification and dilution, fast
PCR amplification for various primer sets and a quicker capillary electrophoresis detection
method, without significant cost increases or instrumentation changes. The methods that were
developed successfully achieved this goal, specifically by reducing processing time by as much
as 37% (using POP-6 detection) and allowing the entire process to be completed in about a
single work day (see Figure 4.1). A second first pass option includes all of the above with
traditional POP-4 detection and takes about 11hr from start to finish (23% reduction). New first
pass options are reproducible and robust, and demonstrated high success rates for buccal
swabs (95-100%), Buccal DNA Collectors™ (85-91%) and old, degraded buccal swabs (86%) that
are comparable to current processes’ success rates (97-99%, 76-90% and 86%, respectively).
These time-related improvements were accompanied by cost savings for reagents,
supplies, analyst labor and instrumentation (see Table 4.1). Reduction in instrument usage can
result in cost savings in different ways. A laboratory could choose to reduce their number of
instruments, thereby reducing maintenance and service contract costs – which can be several
thousand dollars per year, depending on the instrument. Or, a laboratory could choose to
maintain the same number of instruments, but by using them less often, may increase their life
span. This option would also allow increases in sample volume without added instrument costs.
184
Figure 4.1. Comparison of new and current Cellmark Forensics “first pass” processes for databasing reference samples. Through the implementation of normalized extraction (and elimination of quantification and dilution), fast PCR amplification and quicker capillary electrophoresis detection, this new first pass process offers a 37% reduction in processing time.
4.1.1 Normalized Extraction
The basis of the normalized extraction procedure was to reduce bead volume such that
bead binding capacity would be forced to increase to its maximum, thereby resulting in DNA
recoveries that were consistently within a narrow concentration range suitable for fast PCR
amplification. Bead binding capacity was successfully increased using the normalized extraction,
but did not always achieve maximum capacity for every sample. This was not detrimental to the
overall process because the target input DNA concentration range for fast PCR amplification
was wide enough (0.188-3.75ng) to accommodate bead binding capacities less than the
advertised maximum.
1.0 7.0%
0.50 3.5%
2.0 14%
0.75 5.3%
0.50 3.5%
3.3 23%
5.0 35%
1.3 8.8%
Accessioning
Cut
Extraction
Quantification
Dilution
Amplification
Detection
Analysis
1.0 11%
0.50 6%
2.0 22%
1.3 14%
3.0 33%
1.3 14% 0.50
3.5%
9.0hr 14.3hr
New First Pass Procedure Current First Pass Procedure
0.50 3.5%
185
Table 4.1
Time and Cost Savings for New First Pass Processes
Process Reagents Supplies Analyst Labor Instrument
Usage
Normalized Extractiona FLOUstar microplate reader TECAN liquid handler
-$0.22
-$0.03
-0.89hr
Eliminated
-1.0hr Fast PCR Amplification
3μl reaction PowerPlex 16 HSb Identifiler Plusb Identiflerb
6μl reaction Identifilerc
<$0.06
<$0.12
-1hr 4min -1hr 11min -2hr 0min
-2hr 0min
POP-6/22cm Array Detection 3130xl Genetic Analyzer
TBDd
-2hr 0min Note. Changes in processing time (hours) and costs (per sample) are displayed for each step. For fast PCR, reagent costs should increase no more than the cost of KAPA2G; more specific changes to reagent costs are not known at this time, but will be reduced due to the elimination of the need for supplemental AmpliTaq Gold® DNA Polymerase (Identifiler) and the ability to obtain more reactions from the PowerPlex 16 HS and Identifiler Plus kits through complete consumption of their primer sets. aIncludes elimination of quantification and dilution.
bVeriti thermal cycler.
c9700 thermal cycler.
dA small decrease
in polymer costs may occur because a shorter array is used, but is yet to be determined.
As a result, DNA concentrations will be substantially lower using the normalized
extraction compared to that of the current extraction process. Extract quality does not appear
to be compromised using the normalized process; however, the new first pass procedure –
specifically amplification – is highly sensitive to the presence of bead carryover in the sample
extracts. It is imperative that samples remain bead-free, otherwise amplification will suffer and
result in profiles exhibiting PCR inhibition.
Compared to non-normalized extraction, samples processed with the normalized
extraction tended to have higher instances of n-8 stutter and elevated baseline peaks, but not
to the point that the profile would fail from a quality standpoint. Furthermore, during the
186
developmental stages, I attempted to use the same extraction procedure for buccal swabs and
Buccal DNA Collector™ punches, but repeatedly obtained lower DNA recoveries from the punch
samples and/or poorer quality profiles, which resulted in lower first pass rates. Thus, a larger
quantity of beads was ultimately chosen for extraction of buccal punch samples.
Normalized extraction did not require any additional reagents or supplies, though it did
require the use of a different 96-well plate for the elution step on the MPS instrument (one that
was compatible with the MPS and the instrument used to transfer template DNA to the
amplification plate). Factoring cost differences associated with the plate change and
elimination of quantification and dilution, a cost savings of $0.25/sample was achieved.
Additionally, analyst labor and the liquid handler instrument usage were reduced by 0.89hr and
1.0hr, respectively.
4.1.2 Fast PCR Amplification
Overall, low volume fast PCR development using non-fast thermal cyclers was quite
successful and resulted in amplification times of 42-51min for 3μl reactions on a Veriti thermal
cycler. As has been demonstrated previously (Foster & Laurin, 2012; Laurin & Frégeau, 2012),
percent stutter did increase compared to that of standard PCR, but use of a 20% global stutter
filter or modification of the locus specific stutter thresholds in GeneMapper® ID should prevent
excessive stutter peaks from being called by the software. All amplified alleles were concordant
with those obtained using standard PCR. However, it should be noted that a low-level (~100rfu)
artifact was first noted during the validation of Identifiler fast PCR using KAPA2G (without
normalized extraction or alternate detection) and occurred in 1.1% of samples amplified with
187
either a 3μl or 5μl amplification. A post-validation modification was made to the thermal cycling
parameters to increase the annealing temperature to 63°C, which appeared to have solved the
issue, but this artifact did resurface in during validation of the new first pass procedures. Thus,
increasing the annealing temperature appears to have decreased its occurrence, but did not
completely eliminate it. Furthermore, this artifact was not reproducible upon re-amplification
of the same sample and no other non-specific artifacts were observed. Comparatively, Vallone
et al. (2008) reported more frequent occurrences of non-specific product formation for their
10μl Identifiler fast PCR protocol. Allele peak height was slightly reduced using fast PCR, which
that aided in other ways (i.e., allowed a large DNA input range), and was more reproducible
using fast PCR with the KAPA2G™ Fast Multiplex PCR Kit than using standard Identifiler
amplification. Intra- and inter-locus balance were acceptable, though slightly higher inter-locus
imbalance was observed for fast PCR compared to standard.
One key point that made these protocols a success with the normalized extraction
process is their large optimal DNA input range (~0.188-3.75ng), which was larger than the range
I initially determined using non-normalized extracts (0.375-1.50ng). During the assessment of
various fast polymerases, KAPA2G demonstrated the widest range compared to the other
polymerases. Foster and Laurin (2012) and Vallone et al. (2008) developed fast PCR protocols
using the Identifiler primer set with SpeedSTAR and PyroStart/SpeedSTAR, respectively, that
had optimal input ranges of 0.2-1ng and 0.4-1ng DNA, respectively, which is fairly consistent
with the range (0.375-1.00ng) I determined when I assessed SpeedSTAR. However, I do not
think such a small target window would work well with normalized extraction, and would result
in too many under- or over-amplified samples.
188
Development of the various fast PCR amplification protocols was most challenging using
AmpliTaq Gold Fast and SpeedSTAR due to repeated issues with heterozygote peak height
ratios. Giese et al. (2009) and Walsh and Erlich et al. (1992) suggested using a longer
denaturation, but this did not resolve the problem. Use of traditional 3-step PCR cycling –
rather than 2-step – was much more beneficial for these two polymerases (with regard to PHR),
but a 2-step cycle resulted in better PHR than 3-step using Type-it.
Incomplete adenylation (-A) has been shown to often accompany STR profiles obtained
using fast PCR (Giese et al., 2009; Vallone et al., 2008), and was a repeated problem during the
developmental stages of fast protocols in this study, but was easily resolved by increasing final
extension time for nearly all methods. Since I was attempting to keep total amplification to a
minimum, I didn’t want to employ a final extension any longer than absolutely necessary. -A
peak morphology was often weak, but uncalled by GeneMapper® ID because the peak was
below threshold when a 5min (or even 10min for some polymerases) final extension was used.
Thus, most polymerases required a 10-13min final extension to completely eliminate -A. During
the normalized extraction development process, the 6μl Identifiler fast PCR reaction began
demonstrating various performance issues using controls and samples, including -A (10-20min
final extensions) and +A peaks (23min final extension). These were not seen during the fast PCR
development process, which initially lead me to believe such issues were associated with the
normalized extraction procedure and/or buccal punch samples. But since the same problems
were surfacing using the 9947A controls, I could not rule out 9700 performance issues.
Ultimately, the final extension time was increased to 20min for the 6μl Identifiler fast PCR
189
amplification only, in order to best prevent -A. Overall, profile quality of the 6μl Identifiler fast
amplifications was not as good as 3μl fast PCR.
As mentioned above, 2-step PCR cycling did have unacceptable consequences for some
of the fast polymerases (AmpliTaq Gold Fast and SpeedSTAR), but performed very well using
KAPA2G and Type-it. When additional fast PCR protocols were developed for other reaction
volumes and primer sets using KAPA2G, 2-step PCR cycling was successful for all, except with
PowerPlex 16 HS (which also demonstrated an increase in PHR <50%).
Low-level, non-specific amplification was well controlled by adjusting primer annealing
temperature. However, the initial temperature (61°C) I selected for the Identifiler primer set
later proved to be too low and I had to increase it to 63°C. However, despite this change, the
low-level artifact at Amelogenin (~100rfu, ~108bp) initially seen during the Identifiler fast PCR
validation (3μl and 5μl reactions) was seen once more during validation of the normalized
extraction procedure.
Traditionally, most standard amplification protocols utilize a 4°C final hold, but I
assessed a 25°C final hold for all fast PCR protocols because Foster and Laurin (2012) suggested
using 25°C to prevent -A. Using a higher temperature for the final hold did not have any adverse
effects, though I did not specifically see a reduction in -A formation. However, it did result in a
slightly shorter amplification (~1-2min shorter) and demonstrated an increase in peak height
for AmpliTaq Gold Fast.
Fast PCR amplification was developed using the same thermal cyclers (384-well Veriti
and 9700) currently used in the laboratory selected for this project. Many laboratories are
under budgetary constraints and cannot allocate funds for new and/or “fast” thermal cyclers,
190
which are indeed costly. Therefore, I wanted to demonstrate to other laboratories that thermal
cyclers that are commonly used in forensic DNA testing are suitable for fast PCR. Fast thermal
cyclers can achieve shorter amplification times (as low as 19min) than the ones I developed (42-
51min on the Veriti and 51-60min on the 9700), but in my opinion the need for such a short
amplification was not justifiable in this particular laboratory and, therefore, did not warrant
such an expense.
Furthermore, I wanted to develop fast PCR protocols using primer sets that are
commonly used in forensics in conjunction with a reasonably priced, commercially available fast
polymerase. Again, this would demonstrate to other laboratories that there is not a need to
have a specially concocted, home-brewed reaction or significantly increase reaction costs in
order to achieve a shorter amplification process. KAPA2G™ Fast Multiplex PCR Kit is reasonably
priced and resulted in a $0.06/sample cost increase for 3μl amplifications ($0.12 for 6μl); for
laboratories using full 25μl reactions, per sample costs will increase by $0.50, which could be a
substantial amount depending on annual sample throughput. Despite having to purchase a fast
PCR reagent in addition to amplification kit(s), per sample amplification costs may actually
remain the same or slightly decrease compared to the current process utilized at Cellmark
Forensics (yet to be determined). Using the current (standard) amplification process, the
master mix and/or AmpliTaq Gold® DNA polymerase is the limiting reagent in each of the
amplification kits, and implementing these fast PCR protocols will increase the number of
amplification reactions that can be obtained from the Identifiler Plus and PowerPlex 16 HS kits,
while eliminating the need to purchase supplemental AmpliTaq Gold® DNA polymerase for
191
Identifiler. These benefits could offset the cost of KAPA2G, but the extent of which remains to
be determined.
In summary, there are no added costs due to supplies, labor or instrumentation for fast
PCR amplification. Increases to reagent costs may be as high as $0.06/sample for 3μl reactions
(or $0.12 for 6μl), but will likely be less given the elimination of the need for supplemental
AmpliTaq Gold® DNA polymerase and obtaining slightly more reactions from the PowerPlex 16
HS and Identifiler Plus kits compared to standard amplification. Fast PCR amplification results in
a decrease of 1-2hr of instrument usage, depending on the instrument and primer set used (see
Table 4.1).
4.1.3 Alternative POP-6 (22cm Array) Detection
Compared to the traditional POP-4 detection method, the alternative POP-6 method
substantially reduces capillary electrophoresis detection time (24-28min per injection) while
improving precision and resolution for smaller alleles (<200bp), but slightly decreasing ILS sizing
quality and resolution for larger alleles (>200bp). Furthermore, additional migration issues
resulting in miscalled or OL alleles did not occur more frequently using POP-6 than with
traditional detection.
Baseline separation can be corrected by adjusting “voltage number of steps” and
“voltage interval step”. Decreasing run voltage from 15kV to 12-14kV will have a greater
negative impact, as compared to altering run temperature or current stability, on migration
speed, but also results in improved ILS sizing quality. Various combinations of run voltage
(decrease from 15kV) and run temperature (increase from 60°C) further improved ILS sizing
192
quality, precision and resolution. Overall, resolution using the validated POP-6 detection
method is too low for alleles >300bp, as seen by RSL values >1 for alleles of that size, indicating
that one base-pair resolution will not always be achieved for larger amplicons. For this reason, a
new first pass option using traditional POP-4 detection may be the best option, especially for
PowerPlex 16 HS because it has larger-sized alleles than Identifiler Plus and Identifiler. For
PowerPlex 16 HS, five loci have alleles >300bp (D18, Penta E, CSF, Penta D and FGA), compared
to four loci for Identifiler Plus/Identifiler (CSF, D2, D18 and FGA; though the most common FGA
alleles [ 33.2] are <300bp). However, despite reduced resolution, first pass procedures utilizing
POP-6 could still be a viable option if all profiles exhibiting a broad allele peak (as indicated via
GeneMapper® ID) >300bp are reprocessed using POP-4 to confirm the presence of only one
allele at that peak. This is a good check to have in place for traditional POP-4 detection as well,
considering that large-sized alleles differing by one base-pair (e.g., alleles 13.4 and 14 at Penta
D, >423bp) have, on at least one occasion, not been resolved enough to be differentiated by
the analysis software (A. K. LeFebvre, personal communication, November 12, 2014; see Figure
D.15).
Transitioning from the traditional POP-4/36cm array 3130xl setup to POP-6/22cm array
only requires modifications to detection run modules, and results in about a 2hr reduction in
processing time for a full plate (analyst labor is unaffected). Additionally, there is no unit cost
difference between the two polymers or array lengths. However, since the alternative
detection method utilizes a 22cm array, less polymer will be used to fill this shorter array.
However, achieving a cost savings by decreasing polymer usage is in fact more complicated
than might first be considered intuitive. It is possible that no savings will incur as a result of
193
reduced per injection polymer usage because standard 3130xl instrument maintenance
requires that any unused polymer that has been on the instrument for more than the specified
amount of time (one week per Life Technologies, but as dictated by the individual laboratory)
be discarded. Thus, due to this maintenance policy, net polymer usage (and hence cost) may be
the same between the two detection methods when a standard bottle of polymer is not
completely consumed during the allotted time period.
4.2 Future Studies
4.2.1 Resolution
Resolution issues are not new to forensic science or molecular biology. Early DNA
profiling protocols using RFLP (restriction fragment length polymorphism) frequently faced poor
resolution on autoradiographs (or with chemiluminescent detection) of large and/or similarly
sized fragments, coupled with additional complications for high quantity samples (Rudin &
Inman, 2002, pp. 105-114). For example, the D1S7 locus has a 9bp repeat, but its fragment sizes
are 4-12kb, and the technology at the time prevented discrimination of such large fragments
differing by a single repeat (Rudin & Inman, 2002, p. 110). As a result, a continuous allele
system that relied on fragment size binning was used for interpretation purposes, as opposed
to a discrete system, due to the inability to accurately identify number of repeats (Rudin &
Inman, 2002, p. 140). As technology progressed into the era of PCR amplification of STR loci,
discrete allele systems were utilized, and resolution capabilities improved further with the
transition from gel based detection systems (such as the ABI Prism® 373 and 377 DNA
Sequencers; Applied Biosystems) to capillary electrophoresis detection (such as the ABI Prism®
194
310 and ABI 3100/3500/3700 series Genetic Analyzers; Applied Biosystems). Though resolution
improved tremendously compared to that of the earlier RFLP days, the need for greater than
4bp resolution for up to ~300bp (the CODIS loci possess tetranucleotide repeats) was identified
through a population study that revealed the presence of heterozygote alleles differing by 1, 2
or 3bp (Butler et al., 2004). Butler et al. demonstrated POP-6/36cm 1-bp resolution on a 3100
Genetic Analyzer, which was indeed superior to that of POP-4/36cm.
Thus, for forensic STR analysis using the 3130xl, there is indeed a need for 1bp
resolution for alleles as large as 475-500bp (the largest sized alleles currently obtainable using
amplification kits compatible with 3130xl detection). With the alternative POP-6/22cm
detection method developed for this project, 1bp resolution caps out around 300bp, and if
improved upon, this detection process would be more inviting to the forensic community, as
well as other molecular fields utilizing CE detection of similarly sized fragments or alleles. But as
mentioned previously, traditional POP-4/36cm detection has its own resolution limitations, and
has demonstrated unable to provide sufficient resolution between large alleles (>400bp)
differing by a single base-pair. As has been done in the past, scientists must be aware of their
methods’ limitations and work within these confines until technology is adequately improved. It
is well established that 1bp resolution has its limits using the 3130xl with various polymers and
array lengths, but as scientists, we can rely on clues from our data (i.e., the presence of a broad
peak) to suggest the possibility that two peaks are present and investigate further on an as-
needed basis.
195
4.2.2 Integration in Casework
The ultimate goal of this project was to provide a foundation for feasible process
improvement that did not require significant cost increases or instrumentation changes and
that could be applied to other areas, including forensic casework. Fast, high throughput
workflow, at a reasonable cost, is indeed highly desired for forensic DNA laboratories. I have
successfully demonstrated the ability to substantially reduce processing time (by up to 37%)
without additional costs or instrumentation and would be pleased for this project to assist
other laboratories with similar efforts.
Integration into forensic casework would require additional testing on a variety of
poorer quality, forensic type samples (inhibited, degraded, various substrates, mixtures, etc.).
Furthermore, the current FBI QAS guidelines require human-specific quantification prior to
amplification (FBI, 2011), thus a quantification step could not be eliminated while that standard
is still in place. However, normalized extraction could still be beneficial by eliminating the need
for pre-amplification dilution. Such a feat would likely be more challenging given the wide range
of sample qualities and quantities encountered in forensic evidence. However, samples could
be batched based on expected results (confirmed at the quantification step) and processed
with slightly different options based on the batch type (i.e., high quality samples versus low
quantity or degraded samples).
Despite the dynamic nature of forensic DNA technology and other molecular biology
fields, laboratories likely will not benefit from implementing every technological improvement
as it comes along, particularly due to disruptions caused by validation and integration
processes. At the same time, laboratories should not shy away from making their processes
196
more efficient using what they have, rather than feeling that the only way to obtain
improvements sufficient enough to justify the time and costs associated with improving their
process would be to switch to a new chemistry, purchase new/more equipment and/or
instrumentation, etc.
197
APPENDIX A
SUPPLEMENTAL MATERIAL – RUN MODULES USED FOR DEVELOPMENT OF A QUICKER
CAPILLARY ELECTROPHORESIS DETECTION METHOD
198
Table A.1
Default Spectral Calibration Run Modules
Parameter POP-4/36cm POP-4/22cm POP-6/36cm
Oven Temp. 60 60 55 Poly Fill Vol. 6500 5900 6500 Current Stability 5.0 5.0 5.0 PreRun Voltage 15.0 15.0 15.0 PreRun Time 180 180 180 Injection Voltage 1.2 1.0 1.2 Injection Time 18 11 18 Voltage # of Steps 40 10 40 Voltage Step Int. 15 20 15 Data Delay Time 100 1 100 Run Voltage 15 15.0 15.0 Run Time 800 480 1900 Note. Default spectral run modules are listed here as provided by Life Technologies for the various polymer/array length combinations. No default module exists for POP-6/22cm array. Cellmark Forensics uses the default POP-4/36cm array module for dye sets F, G5 and D, except that a 3kV injection is used for dye set D as recommended by Promega.
Table A.2
Spectral Calibration Run Modules Evaluated With Alternative Polymer/Array Combinations
NanoPOP4/36cm POP-6/22cm
Parameter F / G5a F / G5 / D G5 D
Oven Temp. 60 60 60 60 Poly Fill Vol. 6500 6500 5900 5900 Current Stability 5.0 5.0 5.0 5.0 PreRun Voltage 15.0 15.0 15.0 15.0 PreRun Time 180 180 180 180 Injection Voltage 1.2 1.2 1.2 3.0 Injection Time 18 18 18 18 Voltage # of Steps 40 40 40 40 Voltage Step Int. 15 15 15 15 Data Delay Time 100 100 1 1 Run Voltage 15 15 15 15 Run Time 800 500 480 480 Note. Run modules are further differentiated by dye set (F, G5 and D). aThe default POP-4/36cm array spectral run module was tested first for NanoPOP4.
199
Table A.3
Default and Established Run Modules
Default a Established b
POP-4/36cm POP-4/
22cm POP-6/
36cm
POP-4/36cm
Parameter F / G5 D ---c ---c F G5 D
Oven Temp. 60 60 60 55 60 60 60 Poly Fill Vol. 6500 6500 5900 6500 6500 6500 6500 Current Stability 5 5 5 5 5 5 5 PreRun Voltage 15 15 15 15 15 15 15 PreRun Time 180 180 60 180 180 180 180 Injection Voltage 3 3 1 1.2 3 3 3 Injection Time 10 10 22 18 10 10 10 Voltage # of Steps 40 40 10 40 40 40 40 Voltage Step Int. 15 15 20 15 15 15 15 Data Delay Time 1 1 1 240 1 1 1 Run Voltage 15 15 15 15 15 15 15 Run Time 1500 1800 720 2400 1360 1430 1800 Note. Run modules are differentiated by dye set (F, G5 and D). No default run module exists for POP-6/22cm array. aAs provided by Life Technologies or Promega (dye set D only).
bAs used by Cellmark Forensics.
cModule is not used
for a specific dye set.
Table A.4
Run Modules Evaluated With NanoPOP4/36cm Array
F G5 D
Parameter 1 2 3 4 5 1 2 3 1
Oven Temp. 60 60 61 61 60 60 60 60 60 Poly Fill Vol. 6500 6500 6500 6500 6500 6500 6500 6500 6500 Current Stability 5 4 5 5 5 5 5 5 5 PreRun Voltage 15 15 15 15 15 15 15 15 15 PreRun Time 180 180 180 180 180 180 180 180 180 Injection Voltage 3 3 3 3 3 3 3 3 3 Injection Time 10 10 10 10 10 10 10 10 10 Voltage # of Steps 40 40 40 40 40 40 40 40 40 Voltage Step Int. 15 15 15 15 15 15 15 15 15 Data Delay Time 1 1 1 1 1 1 1 1 1 Run Voltage 10 10 10 11 15 15 8 10 15 Run Time 1360 1360 1360 1360 1360 1430 1430 1430 1800 Note. Run modules are differentiated by dye set F, G5 and D.
200
Table A.5
Initial Run Modules Evaluated With POP-6/22cm Array
Parameter 1a 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Oven Temp. 60 60 60 60 60 60 60 60 60 60 60 60 60 63 55 55 55 55 55 55 55 Poly Fill Vol. 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 Current Stability 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 PreRun Voltage 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 PreRun Time 60 180 180 180 180 180 100 60 60 60 60 60 60 60 60 60 60 60 60 60 60 Injection Voltage 1 3 3 3 3 3 3 3 3 3 3 1 3 3 3 3 3 3 3 3 3 Injection Time 22 10 10 10 10 10 10 10 10 10 10 22 10 10 10 10 10 10 10 10 10 Voltage # of Steps 10 40 10 40 30 40 40 10 10 10 10 10 10 10 10 10 40 40 40 40 40 Voltage Step Int. 20 15 20 20 20 15 20 20 20 20 20 20 20 20 20 20 15 15 20 10 10 Data Delay Time 1 1 1 1 1 1 1 400 300 250 200 200 175 175 175 175 200 1 1 1 50 Run Voltage 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 Run Time 720 860 860 860 860 860 860 1000 900 860 860 860 860 860 860 950 1100 1100 1100 1100 1000
Note. Run modules were only tested with dye set G5 for the initial evaluation using Identifiler Plus amplification product. aDefault run module provided by Life Technologies for POP-4/2cm array.
Table A.6
Run Modules Evaluated With POP-6/22cm Array for Peak Height and Pull-up Reduction
Parameter 1 2 3 4 5
Oven Temp. 60 60 60 60 60 Poly Fill Vol. 5900 5900 5900 5900 5900 Current Stability 5 5 5 5 5 PreRun Voltage 15 15 15 15 15 PreRun Time 60 60 60 60 60 Injection Voltage 3 2 2 2 2 Injection Time 7 4 20 10 7 Voltage # of Steps 10 10 10 10 10 Voltage Step Int. 20 20 20 20 20 Data Delay Time 175 175 175 175 175 Run Voltage 15 15 15 15 15 Run Time 860 860 860 860 860 Note. Run modules were only tested with dye set G5 for this evaluation using Identifiler Plus amplification product. Differences between modules are in red. Data was compared to that obtained from traditional POP-4/36cm detection using an established protocol from Cellmark (see G5 in Table A.3), but with a 7sec, rather than 10sec, injection.
201
Table A.7
Run Modules Evaluated With POP-6/22cm Array for ILS Sizing Quality Improvements
Parameter 1 2 3 4 5 6 7 8 9 10 11
Oven Temp. 60 60 60 60 63 63 63 60 60 63 63 Poly Fill Vol. 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 5900 Current Stability 5 5 5 5 5 5 5 4 6 4 6 PreRun Voltage 15 15 15 15 15 15 15 15 15 15 15 PreRun Time 60 60 60 60 60 60 60 60 60 60 60 Injection Voltage 2 2 2 2 2 2 2 2 2 2 2 Injection Time 7 7 7 7 7 7 7 7 7 7 7 Voltage # of Steps 10 10 10 10 10 10 10 10 10 10 10 Voltage Step Int. 20 20 20 20 20 20 20 20 20 20 20 Data Delay Time 1 175 175 175 175 175 175 150 150 150 150 Run Voltage 12 12 13 14 14 13 12 12 12 12 12 Run Time 1000 1500 1500 1300 1300 1170 1170 1170 1200 1200 1200 Note. Run modules were only tested with dye set G5 for this evaluation using Identifiler Plus amplification product. Differences between modules are in red.
Table A.8
Run Modules Evaluated With POP-6/22cm Array for Precision Improvements
Parameter 1 2 3 4 5 6 7
Oven Temp. 60 60 60 63 63 63 63 Poly Fill Vol. 5900 5900 5900 5900 5900 5900 5900 Current Stability 5 5 5 5 5 5 5 PreRun Voltage 12 13 15 12 13 14 15 PreRun Time 60 60 60 60 60 60 60 Injection Voltage 2 2 2 2 2 2 2 Injection Time 7 7 7 7 7 7 7 Voltage # of Steps 10 10 10 10 10 10 10 Voltage Step Int. 20 20 20 20 20 20 20 Data Delay Time 175 175 175 175 175 175 175 Run Voltage 12 13 15 12 13 14 15 Run Time 1180 1050 860 1180 1050 905 860 Note. Run modules were only tested with dye set G5 for this evaluation using Identifiler Plus allelic ladders. Differences between modules are in red. It should be noted that a 60°C 14kV run module was not evaluated because it had previously exhibited slightly lower ILS sizing qualities (see Table 3.2).
202
Table A.9
Validated POP-6/22cm Array Run Modules
Spectral Calibration STR Detection
Parameter 5-Dyea 4-Dyeb 5-Dyea 4-Dyeb
Oven Temp. 63 63 63 63 Poly Fill Vol. 5900 5900 5900 5900 Current Stability 5 5 5 5 PreRun Voltage 14 14 14 14 PreRun Time 180 180 60 60 Injection Voltage 1.2 3 2 2 Injection Time 18 18 7 7 Voltage # of Steps 40 40 10 10 Voltage Step Int. 15 15 20 20 Data Delay Time 1 1 175 175 Run Voltage 14 14 14 14 Run Time 480 480 905 1150 Note. Run modules were validated for Life Technologies 5-dye and Promega 4-dye kits. aLife Technologies kits.
bPromega kits.
203
APPENDIX B
SUPPLEMENTAL MATERIAL – AMPLIFICATION PROTOCOLS USED FOR DEVELOPMENT OF
VARIOUS FAST PCR METHODS
204
Table B.1
Thermal Cycling Parameters for Standard PCR
PCR Step Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex 16 HS (3µl)
Polymerase Activation
95°C 11min 95°C 11min 95°C 11min 95°C 11min 95°C 2min
# of Cycles Denaturation Annealing Extension
26 94°C 1min 59°C 1min 72°C 1min
28 94°C 1min 59°C 1min 72°C 1min
27 94°C 1min 59°C 1min 72°C 1min
26 94°C 20sec 59°C 3min
10 / 18 94°C / 90°C 30seca 60°C 30secb 70°C 45sec
Final Extension 60°C 60min 60°C 60min 60°C 60min 60°C 10min 60°C 30min Hold 4°C 4°C 4°C 4°C 4°C Total Time 2hr 42min 3hr 8min 3hr 0min 2hr 0min 1hr 55min Note. Established protocols used for standard amplification at Cellmark. a29% ramp rate.
b23% ramp rate.
Table B.2
Thermal Cycling Parameters for 3µl Identifiler Fast PCR Development With KAPA2G
2-Step PCR
PCR Step Final Extension Primer Annealing Denaturation Annealing/Extension Initial Activation Final Extension Final
Polymerase Activation
95°C 3min 95°C 3min 95°C 3min 95°C 3min 95°C 1, 2, 3min 95°C 1min 95°C 1min
26 Cycles of: Denaturation Annealing Extension
95°C 15sec 59°C 30sec 72°C 30sec
95°C 15sec 59, 61, 63°C 30sec
95°C 5, 10, 15sec 61°C 30sec
95°C 5sec 61°C 30, 35, 40sec
95°C 5sec 61°C 40sec
95°C 5sec 61°C 40sec
95°C 5sec 61°C 40sec
Final Extension 72°C 1, 2, 5min 72°C 5min 72°C 5min 72°C 5min 72°C 5min 72°C 3, 4, 5, 10min
72°C 10min
Hold 25°C 25°C 25°C 25°C 25°C 25°C 25°C Total Time 50, 51, 54min 41, 40, 39min 36, 38, 40min 36, 38, 40min 38, 39, 40min 36, 37, 38, 43min 43min
Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler.
205
Table B.3
Thermal Cycling Parameters for 3µl Identifiler Fast PCR Development With AmpliTaq Gold Fast
2-Step PCR
PCR Step Final Extension Primer Annealing
Denaturation & Annealing/ Extension
Denaturation Initial Activation Final Hold Annealing/ Extension
Denaturation
Polymerase Activation
95°C 10min 95°C 10min 95°C 10min 95°C 10min 95°C 1, 3, 10min 95°C 10min 95°C 10min 95°C 10min
26 Cycles of: Denaturation Annealing Extension
96°C 5sec 59°C 10sec 68°C 10sec
96°C 5sec 59, 61, 63°C 50sec
96°C 5, 10sec 61°C 40, 50, 60sec
96°C 15, 20sec 61°C 1min
96°C 10sec 61°C 1min
96°C 10sec 61°C 1min
96°C 10sec 61°C 1, 2min
96°C 10sec, 1min 61°C 2min
Final Extension 72°C 1, 5, 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min Hold 4°C 4°C 4°C 4°C 4°C 4, 25°C 25°C 25°C
Total Time 35, 39, 44min 58, 57, 56min 52min, 54min, 56min, 58min, 1hr 1min, 1hr 3min
1hr 5min, 1hr 7min
54min, 55min, 1hr 3min
1hr 3min 1hr 29min 1hr 29min, 1hr 51min
3-Step PCR
PCR Step Standard PCR Annealing & Extension
Denaturation Annealing Extension Final Extension Final
Polymerase Activation
95°C 11min 95°C 10min 95°C 10min 95°C 10min 95°C 10min 95°C 10min 95°C 10min
26 Cycles of: Denaturation Annealing Extension
94°C 1min 59°C 1min 72°C 1min
96°C 1min 61°C 1min 68, 72°C 1min
96°C 10sec, 1min 61°C 1min 68°C 1min
96°C 10sec 61°C 15, 30, 45, 60sec 68°C 1min
96°C 10sec 61°C 45sec 68°C 30, 45, 60sec
96°C 10sec 61°C 45sec 68°C 45sec
96°C 10sec 61°C 45sec 68°C 45sec
Final Extension 60°C 60min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10, 13min 72°C 13min Hold 4°C 25°C 25°C 25°C 25°C 25°C 25°C
Total Time 2hr 42min 1hr 51min 1hr 29min 1hr 10min, 1hr 16min, 1hr 23min, 1hr 29min
1hr 10min, 1hr 16min, 1hr 23min
1hr 16min, 1hr 19min
1hr 19min
Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler. PHR <50% were problematic and resulted testing of more parameters.
206
Table B.4
Thermal Cycling Parameters for 3µl Identifiler Fast PCR Development With SpeedSTAR
2-Step PCR
PCR Step
Primer Annealing
Annealing/Extension Denaturation &
Annealing/Extension
Polymerase Activation
95°C 1min
95°C 1min
95°C 1min
26 Cycles of: Denaturation Annealing Extension
95°C 5sec 59, 61, 63°C 25sec
95°C 5sec 63°C 25, 30, 40sec
95°C 5, 10sec 63°C 25, 30, 40, 50, 60sec
Final Extension 72°C 1min 72°C 1min 72°C 1min Hold 25°C 25°C 25°C Total Time 28, 27min
a 28, 29, 38min 28-49min
3-Step PCR
PCR Step Primer Annealing Extension Annealing Final Extension Primer Annealing Final
Polymerase Activation
95°C 1min 95°C 1min 95°C 1min 95°C 1min 95°C 1min 95°C 1min
26 Cycles of: Denaturation Annealing Extension
98°C 5sec 59, 61, 63°C 15sec 72°C 10sec
98°C 5sec 59°C 15sec 72°C 10, 20sec
98°C 5sec 59°C 15, 20, 25sec 72°C 20sec
98°C 5sec 59°C 25sec 72°C 20sec
98°C 5sec 59, 61, 63°C 25sec 72°C 20sec
98°C 5sec 61°C 25sec 72°C 20sec
Final Extension 72°C 1min 72°C 1min 72°C 1min 72°C 1, 10, 13min 72°C 13min 72°C 13min Hold 25°C 25°C 25°C 25°C 25°C 25°C Total Time 29, 28min
a 29, 33min 33, 38min 38, 47, 50min 50, 49, 48min 49min
Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler. For 2-step PCR cycles, the manufacturer recommends a denaturation temperature of 95°C, but 98°C for 3-step cycling. aFor 61°C and 63°C protocols.
207
Table B.5
Thermal Cycling Parameters for 3µl Identifiler Fast PCR Development With Type-it
3-Step PCR 2-Step PCR
PCR Step Longer Cycles
Longer Cycles Initial Activation Denaturation Annealing/ Extension
Final
Polymerase Activation
95°C 5min
95°C 5min 95°C 1, 2, 5min 95°C 5min 95°C 5min 95°C 5min
26 Cycles of: Denaturation Annealing Extension
95°C 30sec 59°C 45sec 72°C 30sec
95°C 30sec 59°C 1min 15sec
95°C 30sec 59°C 1min 15sec
95°C 10, 20, 30sec 59°C 1min 15sec
95°C 30sec 59°C 55sec, 1min 5sec, 1min 15sec
96°C 30sec 59°C 1min 15sec
Final Extension 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min Hold 25°C 25°C 25°C 25°C 25°C 25°C
Total Time 1hr 13min 1hr 14min 1hr 9min, 1hr 10min, 1hr 14min
1hr 5min, 1hr 9min, 1hr 14min
1hr 5min, 1hr 9min, 1hr 14min
1hr 14min
Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler.
208
Table B.6
Thermal Cycling Parameters for 5µl Identifiler Fast PCR Development With KAPA2G
PCR Step Amplification Mode
# Cycles Annealing/ Extension
# Cycles Final Extension Final
Thermal Cycler Mode 9600 Emulation, Max
Max Max Max Max Max
Polymerase Activation 95°C 1min 95°C 1min 95°C 1min 95°C 1min 95°C 1min 95°C 1min # of Cycles Denaturation Annealing/Extension
26 95°C 5sec 61°C 40sec
26, 28 95°C 5sec 61°C 40sec
26 95°C 5sec 61°C 40, 45, 50sec
26, 27 95°C 5sec 61°C 40sec
27 95°C 5sec 61°C 40sec
26, 27 95°C 5sec 61°C 40sec
Final Extension 72°C 5min 72°C 5min 72°C 5min 72°C 5min 72°C 5, 10min 72°C 10min Hold 25°C 25°C 25°C 25°C 25°C 25°C Total Time 44, 55min 44, 46min 44, 46, 48min 44, 46min 46, 51min 51min Note. Each of the thermal cycling parameters was assessed using a 96-well 9700 thermal cycler (gold or silver block). The final thermal cycling parameters that were developed for the 5µl amplification was also used for 6µl Identifiler fast PCR.
Table B.7
Thermal Cycling Parameters for 3µl Identifiler Plus Fast PCR Development With KAPA2G
PCR Step Final Extension Primer Annealing Initial Activation Denaturation Annealing/ Extension
Final
Polymerase Activation 95°C 3min 95°C 3min 95°C 1, 2, 3min 95°C 1min 95°C 1min 95°C 1min 26 of Cycles: Denaturation Annealing/Extension
95°C 15sec 59°C 1min
95°C 15sec 59, 61, 63°C 1min
95°C 15sec 63°C 1min
95°C 5, 10, 15sec 63°C 1min
95°C 10sec 63°C 50sec, 1min
95°C 10sec 63°C 50sec
Final Extension 72°C 1, 5, 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min Hold 25°C 25°C 25°C 25°C 25°C 25°C Total Time 50, 55, 59min 59, 58, 57min 55, 56, 57min 51, 53, 55min 49, 53min 49min Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler.
209
Table B.8
Thermal Cycling Parameters for 3µl PowerPlex 16 HS Fast PCR Development With KAPA2G
3-Step PCR 2-Step PCR 3-Step PCR
PCR Step Final Extension Primer Annealing
Ramp Rates Annealing Extension Denaturation Initial Activation
Final Hold Final
Polymerase Activation
96°C 2min 96°C 2min
96°C 2min 96°C 2min 96°C 2min 96°C 2min 96°C 1, 2min
96°C 1min 96°C 1min
10 of Cycles: Denaturation Annealing Extension
94°C 15sec 60°C 30sec
a
70°C 30secc
94°C 15sec 58, 60, 62°C 1min
a
94°C 15sec 60°C 30sec
a,b
70°C 30secb,c
94°C 15sec 60°C 15, 30sec 70°C 30sec
94°C 15sec 60°C 15sec 70°C 15, 20, 30sec
94°C 5, 10, 15sec 60°C 15sec 70°C 30sec
94°C 15sec 60°C 15sec 70°C 30sec
94°C 15sec 60°C 15sec 70°C 30sec
94°C 15sec 60°C 15sec 70°C 30sec
18 of Cycles: Denaturation Annealing Extension
90°C 15sec 60°C 30sec
a
70°C 30secc
90°C 15sec 58, 60, 62°C 1min
a
90°C 15sec 60°C 30sec
a,b
70°C 30secb,c
90°C 15sec 60°C 15, 30sec 70°C 30sec
90°C 15sec 60°C 15sec 70°C 15, 20, 30sec
90°C 5, 10, 15sec 60°C 15sec 70°C 30sec
90°C 15sec 60°C 15sec 70°C 30sec
90°C 15sec 60°C 15sec 70°C 30sec
90°C 15sec 60°C 15sec 70°C 30sec
Final Extension 72°C 1, 5, 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min 72°C 10min Hold 4°C 4°C 4°C 4°C 4°C 4°C 4°C 4, 25°C 25°C
Total Time 1hr 12min, 1hr 16min, 1hr 21min
1hr 12min, 1hr 10min, 1hr 8min
1hr 0min, 1hr 21min
53min, 1hr 0min
45, 48, 53min 48, 50, 53min 52, 53min 52, 51min 51min
Note. Each of the thermal cycling parameters was assessed using a 384-well Veriti thermal cycler. All ramp rates are 100%, unless otherwise noted. a29% ramp rate.
balso tested at 100% ramp rate.
c23% ramp rate.
210
Table B.9
Validated Thermal Cycling Parameters for 3µl, 5µl and 6µl Fast PCR Protocols With KAPA2G
PCR Step Identifiler (3µl) Identifiler (5µl) Identifiler (6µl) Identifiler Plus (3µl) PowerPlex 16 HS (3µl)
Polymerase Activation
95°C 1min 95°C 1min 95°C 1min 95°C 1min 96°C 1min
# of Cycles Denaturation Annealing Extension
26 95°C 5sec 63°C 40sec
27 95°C 5sec 61°C 40sec
27 95°C 5sec 63°C 40sec
26 95°C 10sec 63°C 50sec
10 / 18 94°C / 90°C 15sec 60°C 15sec 70°C 30sec
Final Extension 72°C 10min 72°C 10min 72°C 20min 72°C 10min 72°C 10min Hold 25°C 25°C 25°C 25°C 25°C Total Time 42min 51min 1hr 0min 49min 51min Note. Final, validated fast PCR protocols are displayed above, including post-validation modifications for 3µl and 6µl Identifiler. All 3µl amplifications are performed on 384-well Veriti thermal cyclers, and 5µl or 6µl amplifications are performed on 96-well gold or silver block 9700 thermal cyclers using the maximum ramp rate mode.
211
Figure B.1. Three models for preferential amplification based on locus size. Model A assumes that no locus will amplify the same as another and that amplification will decrease at an equal rate from the smallest to largest locus. Model B categorizes the loci into four groups of four loci (based on size) and assumes that the four smallest loci will amplify the same as each other, followed by equal rates of decreases in amplification for the remaining three groups of four loci. Model C categorizes the loci into nine groups of similarly sized loci and assumes that similarly sized loci will amplify the same as each other, with equal rates of decreases in amplification from the smallest to the largest group.
y = -320.0x + 6320 R² = 1.000
0
2000
4000
6000
8000
LPH
(rf
u)
Model A
y = -376.5x + 6800 R² = 0.9412
0
2000
4000
6000
8000
LPH
(rf
u)
Model B
y = -334.4x + 6649 R² = 0.9826
0
2000
4000
6000
8000
LPH
(rf
u)
Model C
212
Figure B.2. CPALS values for three models of preferential amplification. The resulting CPALS ranges obtained from the three models can be used to describe the relationship between the LPH of the smallest locus (LPHS) to that of the largest locus (LPHL) in an effort to identify the degree of preferential amplification based on locus size. When loci are arranged in ascending order based on size, a positive CPALS indicates preferential amplification of smaller loci, whereas a negative value indicates preferential amplification of larger loci.
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0 2 4 6 8 10 12
CP
ALS
LPHS:LPHL
Model A Model B Model C
213
Table B.10
CPALS Values to Describe Preferential Amplification Based on Locus Size
LPHS:LPHL CPALS Range
1.25 0.01316 - 0.01518 1.50 0.02174 - 0.02477 1.75 0.02778 - 0.03137 2.00 0.03226 - 0.03620 2.25 0.03571 - 0.03988 2.50 0.03846 - 0.04278 2.75 0.04070 - 0.04513 3.00 0.04249 - 0.04706 3.25 0.04412 - 0.04868 3.50 0.04545 - 0.05006 3.75 0.04661 - 0.05125 4.00 0.04762 - 0.05229 4.25 0.04851 - 0.05320 4.50 0.04930 - 0.05400 4.75 0.05000 - 0.05472 5.00 0.05063 - 0.05536 5.25 0.05120 - 0.05594 5.50 0.05172 - 0.05647 5.75 0.05220 - 0.05695 10.0 0.05660 - 0.06138
Note. CPALS values that are calculated from a linear regression line (negative slope/y-intercept) from LPH values plotted per locus (arranged in ascending order based on size) can be used to describe the ratio of the smallest locus’ sum of peak heights (LPHS) to the largest locus’ sum of peak heights (LPHL). For example, a CPSLS of 0.05000 would indicate that the smallest locus’ (Amelogenin) sum of peak heights is 3.75-5.25x greater than the largest locus’ (CSF) sum of peak heights for a particular Identifiler profile. The strength of this relationship is only valid if the plot’s linear regression line has a moderate to strong (>0.3) R
2.
214
APPENDIX C
SUPPLEMENTAL MATERIAL – PARAMETERS FOR NORMALIZED EXTRACTION DEVELOPMENT
215
Table C.1
Evaluated Parameters for Normalized Extraction Development
Swabs Punches
Extraction Parameter
Initial Evaluation
Large Data Set Initial Evaluation
Wash Buffer Volume
Large Data Set Incubation Length
Proteinase K Volume
Post-Incubation Vortex
Large Data Set
Lysis Buffer (µl) 300 300 300 300 300 300 300 300 300 Proteinase K (µl) 5.0 5.0 5.0 5.0 5.0 5.0 5.0, 10 5.0 5.0 Pre-Incubation Vortex
Yes Yes Yes Yes Yes Yes Yes Yes Yes
Sonication 2min 2min 2min 2min 2min 2min 2min 2min 2min 56°C Incubation 1.5hr 1.5hr 1.5hr 1.5hr 1.5hr 1.5hr / overnight 1.5hr 1.5hr 1.5hr Post-Incubation Vortex
No No No No No No No Yes, No No
Purification Buffer:Bead Mixture (µl)
100:5, 100:2, 100:1, 100:0.5, 100:0.25 100µl
100:1, 100:0.5 100µl
100:5, 100:2, 100:1, 100:0.5, 100:0.25 100µl
100:2, 100:1 100µl
100:0.5 100µl
100:0.5 100µl
100:0.5 100µl
100:1, 100:0.5 100µl
100:1 100µl
Wash Buffer 1 / 2 (µl)
125 / 125 125 / 125 125 / 125 125 / 125 250 / 125 250 / 250
125 / 125 125 / 125 125 / 125 125 / 125 125 / 125
Elution Buffer (µl) 60 and 80 60 120 120 60 60 60 80, 60 80 DNA Input (µl) 3µl Fasta 3µl Standarda 6µl Fastc 6µl Standardc
0.9b 1.2b 0.9, 1.8d 0.9, 1.8d
0.9 0.9
1.0, 1.8d 1.0, 2.4d
0.7, 1, 1.8 0.7, 1, 1.8, 2.4
0.7, 1, 1.8 0.7, 1, 1.8, 2.4
0.5, 1, 1.8 0.5
1
1, 1.8
1
Detection POP-4 POP-6
3kV 7, 8, 10sec
3kV 4-12sec 2kV 4-12sec
3kV 7, 10sec
3kV 7, 10sec
3kV 7, 10sec 2kV 5-12sec 2kV 5-12sec
3kV 10sec
3kV 10sec
3kV 10sec 2kV 5, 7sec
2kV 5-12sec 2kV 5-12sec
Note. Various parameters were tested in order to achieve normalized extractions suitable for fast PCR. aIdentifiler, Identifiler Plus and PowerPlex 16 HS.
bOnly 0.25µl bead/60µl elution and 0.50µl bead/80µl elution samples were amplified.
bOnly 0.25µl bead/60µl
elution samples were amplified in order to assess if 6µl Identifiler amplification issues were isolated to punch samples only. cIdentifiler.
dOnly 0.5µl and 1.0µl
bead samples were amplified.
216
Table C.2
Validated Normalized ChargeSwitch Extraction
Extraction Parameter Buccal Swabs Buccal DNA Collector™ Punches
Lysis Buffer (µl) 300 300 Proteinase K (µl) 5.0 5.0 Pre-Incubation Vortex Yes Yes Sonication 2min 2min 56°C Incubation 1.5hr 1.5hr Post-Incubation Vortex No No Purification Buffer:Bead Mixture (µl)
100:0.5 100µl
100:1 100µl
Wash Buffer 1 / 2 (µl) 125 / 125 125 / 125 Elution Buffer (µl) 60 80 DNA Input (µl) 3µl Fasta 6µl Fastb
0.9
0.9 1
Detection POP-4 POP-6
3kV 7sec 2kV 7sec
2kV 7sec 2kV 7sec
Note. The validated extraction parameters for swabs and punches that result in normalized extracts suitable for fast PCR are listed above. aIdentifiler, Identifiler Plus and PowerPlex 16 HS.
bIdentifiler.
217
APPENDIX D
SUPPLEMENTAL MATERIAL – RESULTS
218
Figure D.1. Precision of allele sizing for traditional and alternative detection methods. Allele sizing data was obtained from 12 Identifiler Plus allelic ladders for traditional POP-4/36cm and seven alternative POP-6/22cm detection methods (n=2460 alleles per method). Standard deviation of allele sizing increased as allele size increased for all detection methods. Lines of best fit are displayed for each, including equations and R2 values.
60 C 12kV: y = 0.0004x + 0.0011; R² = 0.4295
60°C 13kV: y = 0.0002x + 0.0189; R² = 0.3285
60°C 15kV: y = 0.0004x + 0.0228; R² = 0.3493
63°C 12kV: y = 0.0003x + 0.0271; R² = 0.3063 63°C 13kV: y = 0.0002x + 0.0286; R² = 0.2119
63°C 14kV: y = 0.0002x + 0.0342; R² = 0.2361
63°C 15kV: y = 0.0003x + 0.0277; R² = 0.3631
POP-4: y = 0.0003x + 0.0024; R² = 0.5229
0.00
0.04
0.08
0.12
0.16
0.20
0 50 100 150 200 250 300 350 400
Stan
dar
d D
evia
tio
n
Allele Size (bp)
219
Figure D.2. One base-pair resolution using traditional and alternative detection methods. Valley value (Vv) increases as allele size increases (left to right) and resolution decreases. POP-6 resolution declines at a faster rate for alleles >200bp (right) compared to POP-4. These samples were obtained from either PowerPlex 16 (left and right) or Identifiler Plus (middle) amplification product detected using traditional POP-4/36cm detection or alternative POP-6/22cm (validated) detection. Allele call, allele size (bp) and allele height (rfu) are indicated for each allele.
POP-4: VV=0.3513
POP-6: VV=0.3759
POP-4: VV=0.4033
POP-6: VV=0.3588
POP-4: VV=0.4556
POP-6: VV=0.5586
220
A) POP-4
B) POP-6
Figure D.3. Representative Identifiler Plus allelic ladders from traditional and alternative detection methods. Allele calls and allele size (bp) are displayed for each allele.
221
A) POP-4
B) POP-6
Figure D.4. Representative PowerPlex 16 allelic ladders from traditional and alternative detection methods. Allele calls and allele size (bp) are displayed for each allele.
222
A) POP-4 B) POP-6
Figure D.5. Representative Identifiler Plus profiles from traditional and alternative detection methods. Profiles were obtained through standard amplification. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele.
A) POP-4 B) POP-6
Figure D.6. Representative PowerPlex 16 profiles from traditional and alternative detection methods. Profiles were obtained through standard amplification. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele.
223
A) AmpliTaq Gold Fast B) KAPA2G C) SpeedSTAR
D) Type-it E) Standard PCR
Figure D.7. Representative 3μl Identifiler profiles from fast and standard PCR. Profiles were obtained using each of the optimized fast PCR methods (A-D) and standard PCR (E) with traditional POP-4 detection. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele.
224
A) 3μl Identifiler Fast PCR B) 5μl Identifiler Fast PCR
C) 6μl Identifiler Fast PCR D) 6μl Identifiler Standard PCR
Figure D.8. Representative Identifiler profiles from fast and standard PCR. Profiles were obtained using each of the validated Identifiler/KAPA2G fast PCR reaction volumes (A-C) and 6μl Identifiler standard PCR (D) with traditional POP-4 detection. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele.
225
A) 3μl Identifiler Plus Fast PCR B) 3μl Identifiler Plus Standard PCR
Figure D.9. Representative Identifiler Plus profiles from fast and standard PCR. Profiles were obtained using the validated Identifiler Plus/KAPA2G fast PCR method (A) and Identifiler Plus standard PCR (B) with traditional POP-4 detection. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele. A) 3μl PowerPlex 16 HS Fast PCR B) 3μl PowerPlex 16 HS Standard PCR
Figure D.10. Representative PowerPlex 16 HS profiles from fast and standard PCR. Profiles were obtained using the validated PowerPlex 16 HS/KAPA2G fast PCR method with traditional POP-4 detection (A) and PowerPlex 16 HS standard PCR (B) with alternative POP-6 detection. Allele calls, allele size (bp) and peak height (rfu) are displayed for each allele.
226
New First Pass: Swabs 3µl PowerPlex 16 HS Fast (POP-4)
New First Pass: Swabs 3µl PowerPlex 16 HS Fast (POP-6)
Current First Pass: Swabs 3µl PowerPlex 16 HS
New First Pass: Punches 3µl PowerPlex 16 HS Fast (POP-4)
New First Pass: Punches 3µl PowerPlex 16 HS Fast (POP-6)
Current First Pass: Punches 3µl PowerPlex 16 HS
Figure D.11. Representative profiles obtained from new and current 3µl PowerPlex 16 HS first pass options. Profiles obtained from swabs are on the left and punches are on the right.
227
New First Pass: Swabs 3µl Identifiler Plus Fast (POP-4)
New First Pass: Swabs 3µl Identifiler Plus Fast (POP-6)
Current First Pass: Swabs 3µl Identifiler Plus
New First Pass: Punches 3µl Identifiler Plus Fast (POP-4)
New First Pass: Punches 3µl Identifiler Plus Fast (POP-6)
Current First Pass: Punches 3µl Identifiler Plus
Figure D.12. Representative profiles obtained from new and current 3µl Identifiler Plus first pass options. Profiles obtained from swabs are on the left and punches are on the right.
228
New First Pass: Swabs 3µl Identifiler Fast (POP-4)
New First Pass: Swabs 3µl Identifiler Fast (POP-6)
Current First Pass: Swabs 3µl Identifiler
New First Pass: Punches 3µl Identifiler Fast (POP-4)
New First Pass: Punches 3µl Identifiler Fast (POP-6)
Current First Pass: Punches 3µl Identifiler
Figure D.13. Representative profiles obtained from new and current 3µl Identifiler first pass options. Profiles obtained from swabs are on the left and punches are on the right.
229
New First Pass: Punches 6µl Identifiler Fast (POP-4)
New First Pass: Punches 6µl Identifiler Fast (POP-6)
Current First Pass: Punches 6µl Identifiler
Figure D.14. Representative profiles obtained from new and current 6µl Identifiler Plus first pass options. Profiles are from punches only.
230
Figure D.15. One base-pair resolution for large sized alleles using traditional detection. Poor resolution was obtained between alleles 13.4 (“OL”) and 14 at the Penta D locus of a PowerPlex 16 HS profile obtained using traditional POP-4/36cm detection, such that both alleles could not be detected by GeneMapper® ID from the same injection. The Broad Peak (“BD”) quality flag was marked by a yellow triangle in the first and subsequent injections, indicating that a peak broader than what would be expected for a single allele was present. This signaled the analyst to re-process the sample. Though both alleles could never be detected by the software during the same injection (not all data shown), multiple injections did provide allele size (bp) and peak height (rfu) information for both alleles.
231
REFERENCES
Aebischer, A., Beer, M., & Hoffman, B. (2014, August). Development and validation of rapid
magnetic particle based extraction protocols. Virology Journal, 11(1), 137.
Anslinger, K., Bayer, B., Rolf, B., Keil, W., & Eisenmenger, W. (2005, May). Application of the
BioRobot EZ1 in a forensic laboratory. Legal Medicine, 7(3), 164-168.
Applied Biosystems. (2004, November). GeneMapper® ID Software Versions 3.1 and 3.2 -
Human Identification Analysis: Tutorial, Part Number 4357520 Rev. A. Foster City, CA:
Applied Biosystems.
Applied Biosystems. (2006a, August). AmpFℓSTR® Identifiler® PCR Amplification Kit: User's
Manual, Part Number 4323291 Rev. D. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2006b, April). Quantifiler™ Kits - Quantifiler™ Human DNA Quantification
Kit and Quantifiler™ Y Human Male DNA Quantification Kit: User's Manual, Part Number
4344790 Rev. D. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2007, January). Applied Biosystems 3130/3130xl Genetic Analyzers:
Getting Started Guide, Part Number 4352715 Rev. C. Foster City, CA: Applied
Biosystems.
Applied Biosystems. (2009, October). AmpFℓSTR® Identifiler® Plus PCR Amplification Kit: User's
Guide, Part Number 4440211 Rev. A. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2010a, June). AmpliTaq Gold® DNA Polymerase, LD: Protocol, Part Number
4338858, Rev. D. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2010b, June). AmpliTaq Gold® Fast PCR Master Mix, UP (2X): Protocol, Part
Number 4391138 Rev. C. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2010c, June). Applied Biosystems Veriti™ Thermal Cycler: User Guide, Part
Number 4375799 Rev. E. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2010d, August). Product Brochure Veriti® Thermal Cycler, Publication
104BR05-05. Foster City, CA: Applied Biosystems.
Applied Biosystems. (2010e, August). Specification Sheet: Veriti™ Thermal Cyclers and
GeneAmp® PCR System 9700, Publication 104SP08-01. Foster City, CA: Applied
Biosystems.
232
Azco Biotech, Inc. (2012, December). Sequencing Polymers. Retrieved from Azco Biotech:
http://www.azcobiotech.com/reagents/NPPolymer.php
Bio-Rad Laboratories, Inc. (2008). C1000™ Thermal Cycler: Instruction Manual, 10010305 Rev B.
Mississauga, ON, Canada: Bio-Rad Laboratories, Inc.
Buel, E., LaFountain, M., Schwartz, M., & Walkinshaw, M. (2001, March). Evaluation of capillary
electrophoresis performance through resolution measurements. Journal of Forensic
Sciences, 46(2), 341-345.
Butler, J. M. (2012). Advanced topics in forensic DNA typing: Methodology. Waltham, MA:
Academic Press.
Butler, J. M., Buel, E., Crivellente, F., & McCord, B. R. (2004). Forensic DNA typing by capillary
electrophoresis using the ABI Prism 310 and 3100 genetic analyzers for STR analysis.
Electrophoresis, 25, 1397-1412.
Daniel, R., Santos, C., Phillips, C., Fondevila, M., van Oorschot, R., Carracedo, Á., Lareu, M. V.,
McNevin, D. (2015, January). A SNaPshot of next generation sequencing for forensic SNP
analysis. Forensic Sci Int Genet, 8(1), 50-60.
Davis, C. P., King, J. L., Budowle, B., Eisenberg, A. J., & Turnbough, M. A. (2012, January).
Extraction platform evaluations: A comparison of Automate Express™, EZ1® Advanced
XL, and Maxwell® 16 Bench-top DNA extraction systems. Legal Medicine, 14(1), 36-39.
Eppendorf. (2005). Mastercycler® ep, B 5341 900.060-03/1003. Hamburg, Germany: Eppendorf
AG.
FBI. (2011, September). Quality assurance standards for forensic DNA testing laboratories,
effective 9-1-2011. Retrieved from The Federal Bureau of Investigation:
http://www.fbi.gov/about-us/lab/biometric-analysis/codis/qas-standards-for-forensic-
dna-testing-laboratories-effective-9-1-2011
Fordyce, S. L., Mogensen, H. S., Børsting, C., Lagacé, R. E., Chang, C.-W., Rajagopalan, N., &
Morling, N. (2015, January). Second-generation sequencing of forensic STRs using the
Ion Torrent™ HID STR 10-plex and the Ion PGM™. Forensic Sci Int Genet, 8(1), 132-140.
Foster, A., & Laurin, N. (2012, March). Development of a fast PCR protocol enabling rapid
generation of AmpFℓSTR® Identifiler®profiles for genotyping of human DNA.
Investigative Genetics, 1-11.
Frantz Burger, M., Song, E. Y., & Schumm, J. W. (2005, August). Buccal DNA samples for DNA
typing: New collection and processing methods. BioTechniques, 39(2), 257-261.
233
Frégeau, C. J., & De Moors, A. (2012, September). Competition for DNA binding sites using
Promega DNA IQ™ paramagnetic beads. Forensic Sci Int Genet, 6(5), 511-522.
Frégeau, C. J., Lett, M. C., & Fourney, R. M. (2010, October). Validation of a DNA IQ™-based
extraction method for TECAN robotic liquid handling workstations for processing
casework. Forensic Sci Int Genet, 4(5), 292-304.
Gangano, S., Elliot, K., Anoruo, K., Gass, J., & Buscaino, J. (2013). DNA investigative lead
development from blood and saliva samples in less than two hours using the RapidHIT™
Human DNA Identification System. Forensic Sci Int Genet Suppl Ser, 4(1), e43-e44.
Ge, J., Eisenberg, A., & Budowle, B. (2012, January). Developing criteria and data to determine
best options for expanding the core CODIS loci. Investigative Genetics, 3(1), 1-14.
Giese, H., Lam, R., Selden, R., & Tan, E. (2009, November). Fast multiplexed polymerase chain
reaction for conventional and microfluidic short tandem repeat analysis. Journal of
Forensic Sciences, 54(6), 1287-1296.
Gill, P., Puch-Solis, R., & Curran, J. (2009, March). The low-template-DNA (stochastic) threshold
– its determination relative to risk analysis for national DNA databases. Forensic Sci Int
Genet, 3(2), 104-111.
Gray, K., Crowle, D., & Scott, P. (2014, September). Direct amplification of casework bloodstains
using the Promega PowerPlex® 21 PCR Amplification System. Forensic Sci Int Genet, 12,
86-92.
Greenspoon, S. A., Ban, J. D., Sykes, K., Ballard, E. J., Edler, S. S., Baisden, M., & Covington, B. L.
(2004, January). Application of the BioMek® 2000 Laboratory Automation Workstation
and the DNA IQ™ System to the extraction of forensic casework samples. Journal of
Forensic Sciences, 49(1), 29-39.
Halsall, A., Ravetto, P., Reyes, Y., Thelwell, N., Davidson, A., Gaut, R., & Little, S. (2008,
February). The quality of DNA extracted from liquid or dried blood is not adversely
affected by storage at 4°C for up to 24 h. International Journal of Epidemiology, 37, i7-
i10.
Hansen, T. v., Simonsen, M. K., Nielsen, F. C., & Hundrup, Y. A. (2007, October). Collection of
blood, saliva, and buccal cell samples in a pilot study on the Danish Nurse Cohort:
Comparison of the response rate and quality of genomic DNA. Cancer Epidemiology,
Biomarkers & Prevention, 16, 2072-2076.
234
Hedman, J., Albinsson, L., Ansell, C., Tapper, H., Hansson, O., Holgersson, S., & Ansell, R. (2008,
June). A fast analysis system for forensic DNA reference samples. Forensic Sci Int Genet,
2(3), 184-189.
Heller, C. (1999). Separation of double-stranded and single-stranded DNA in polymer solutions:
II. Separation, peak width and resolution. Electrophoresis, 20(10), 1978-1986.
Hoogendoorn, B., Norton, N., Kirov, G., Williams, N., Hamshere, M. L., Spurlock, G., Austin, J.,
Stephens, M. K., Buckland, P. R., Owen, M. J., O'Donovan, M. C. (2000). Cheap, accurate
and rapid allele frequency estimation of single nucleotide polymorphisms by primer
extension and DHPLC in DNA pools. Human Genetics, 107, 488-493.
Invitrogen™ Life Technologies. (2005, January). Instruction Manual: ChargeSwitch® Forensic
DNA Purification Kits, Version A. Carlsbad, CA: Invitrogen Corporation.
Kapa Biosystems. (2014a, January). Technical Data Sheet: KAPA2G Fast Multiplex PCR Kit,
KR0378 – v2.14. Boston, MA: Kapa Biosystems.
Kapa Biosystems. (2014b, November). KAPA2G Fast Multiplex. Retrieved from Kapa Biosystems:
http://www.kapabiosystems.com/product-applications/products/pcr-2/kapa2g-fast-
multiplex-pcr-kits/
Kopf-Sill, A. K. (2001). Patent No. US6303343 B1. USA.
Koumi, P., Green, H. E., Jordan, D., Lahec, S., Livett, R. J., Tsang, K. W., & Ward, D. M. (2004).
Evaluation and validation of the ABI 3700, ABI 3100, and the MegaBACE 1000 capillary
array electrophoresis instruments for use with short tandem repeat microsatellite
typing in a forensic environment. Electrophoresis, 25, 2227-2241.
LaRue, B. L., Moore, A., King, J. L., Marshall, P. L., & Budowle, B. (2014, November). An
evaluation of the RapidHIT® system for reliably genotyping reference samples. Forensic
Sci Int Genet, 13, 104-111.
Laurin, N., & Frégeau, C. (2012, January). Optimization and validation of a fast amplification
protocol for AmpFℓSTR® Profiler Plus® for rapid forensic human identification. Forensic
Sci Int Genet, 6(1), 47-57.
Leclair, B., Sgueglia, J. B., Wojtowicz, P. C., Juston, A. C., Frégeau, C. J., & Fourney, R. M. (2003,
September). STR DNA typing: Increased sensitivity and efficient sample consumption
using reduced PCR reaction volumes. Journal of Forensic Sciences, 48(5), 1001-1013.
235
Lennard Richard, M. L., Harper, K. A., Craig, R. L., Onorato, A. J., Robertson, J. M., & Donfack, J.
(2012, July). Evaluation of mRNA marker specificity for the identification of five human
body fluids by capillary electrophoresis. Forensic Sci Int Genet, 6(4), 452-460.
Life Technologies. (2012, June). User Guide: AmpFℓSTR® Identifiler® Direct PCR, Publication Part
Number 4415125 Rev. H. Carlsbad, CA: Life Technologies.
Life Technologies. (2014a, November). AmpliTaq Gold® Fast PCR Master Mix. Retrieved from
Life Technologies:
(http://www.lifetechnologies.com/order/catalog/product/4390939?CID=search-product
Life Technologies. (2014b, September). User Guide: GlobalFiler® PCR Amplification Kit,
Publication Part Number 4477604 Rev. C. Calsbad, CA: Life Technologies.
Life Technologies. (2014c, November 11). PCR Thermal Cyclers Support - Getting Started.
Retrieved from Life Technologies:
http://www.lifetechnologies.com/us/en/home/technical-resources/technical-reference-
library/pcr-cdna-synthesis-support-center/pcr-thermal-cyclers-support/pcr-thermal-
cyclers-support-getting-started.html
Life Technologies. (2014d, November 10). POP-4® Polymer for 3130/3130xl Genetic Analyzers.
Retrieved from Life Technologies:
http://www.lifetechnologies.com/order/catalog/product/4363752?ICID=search-product
Life Technologies. (2014e, November 10). POP-6™ Polymer for 3130/3130xl Genetic Analyzers.
Retrieved from Life Technologies:
https://www.lifetechnologies.com/order/catalog/product/4363783?ICID=search-
product
Life Technologies. (2014f, November 10). POP-7™ Polymer for 3130/3130xl Genetic Analyzers.
Retrieved from Life Technologies:
https://www.lifetechnologies.com/order/catalog/product/4363785?ICID=search-
product
Lowe, A., Murray, C., Whitaker, J., Tully, G., & Gill, P. (2002, September). The propensity of
individuals to deposit DNA and secondary transfer of low level DNA from individuals to
inert surfaces. Forensic Sci Int Genet, 129(1), 25-34.
Moretti, T. R., Baumstark, A. L., Defenbaugh, D. A., Keys, K. M., Brown, A. L., & Budowle, B.
(2001, May). Validation of STR typing by capillary electrophoresis. Journal of Forensic
Sciences, 46(3), 661-676.
236
Myers, B. A., King, J. L., & Budowle, B. (2012, September). Evaluation and comparative analysis
of direct amplification of STRs using PowerPlex® 18D and Identifiler® Direct systems.
Forensic Sci Int Genet, 6(5), 640-645.
New England BioLabs® Inc. (2014, November 9). Fast PCR. Retrieved from NEB:
https://www.neb.com/applications/dna-amplification-and-pcr/specialty-pcr/fast-pcr
Park, S. J., Kim, J. Y., Yang, Y. G., & Lee, S. H. (2008, March). Direct STR amplification from whole
blood and blood- or saliva-spotted FTA DNA purification. Journal of Forensic Sciences,
53(2), 335-341.
Parsons, L., & Bright, J.-A. (2012, December). A manual and automated method for the forensic
analysis of DNA from buccal samples on Whatman Indicating FTA Elute Cards. Australian
Journal of Forensic Sciences, 44(4), 393-402.
Phillips, K., McCallum, N., & Welch, L. (2012, March). A comparison of methods for forensic DNA
extraction: Chelex-100® and the QIAGEN DNA Investigator Kit (manual and automated).
Forensic Sci Int Genet, 6(2), 282-285.
Promega. (2012, October). Technical Manual: PowerPlex® Fusion System, Part# TMD039.
Madison, WI: Promega.
Promega. (2013, November). Techincal Bulletin: DNA IQ™ System—Database Protocol, TB297.
Madison, WI: Promega.
Promega. (2014a, May). Technical Manual: PowerPlex® 16 HS System, Part# TMD022. Madison,
WI: Promega.
Promega. (2014b, May). Technical Manual: PowerPlex® 18D System, Part# TMD031. Madison,
WI: Promega.
Promega. (2014c, July). Technical Manual: PowerPlex® 21 System, Part# TMD034. Madison, WI:
Promega.
Puritan®. (2013, December 3). Forensic swab adhesive change: A Puritan customer case study.
Retrieved from Puritan: http://blog.puritanmedproducts.com/bid/351275/Forensic-
Swab-Adhesive-Change-A-Puritan-Customer-Case-Study
QIAGEN. (2009a, April). EZ1® DNA Investigator Handbook, Fourth Edition. Valencia, CA:
QIAGEN.
QIAGEN. (2009b, December). Type-it® Microsatellite PCR Kit Handbook. Valencia, CA: QIAGEN.
237
QIAGEN. (2014, November). http://www.qiagen.com/us/products/catalog/assay-
technologies/end-point-pcr-and-rt-pcr-reagents/type-it-microsatellite-pcr-kit. Retrieved
from QIAGEN: http://www.qiagen.com/us/search.aspx?q=type-
it%20microsatellite%20pcr%20kit#&&pg=1
Richards, B., Skoletsky, J., Shuber, A. P., Balfour, R., Stern, R. C., Dorkin, H. L., Parad, R. B., Witt,
D., Klinger, K. W. (1993, February). Multiplex PCR amplification from the CFTR gene
using DNA prepared from buccal brushes/swabs. Human Molecular Genetics, 2(2), 159-
163.
Rudin, N., & Inman, K. (2002). An introduction to forensic DNA analysis, Second Edition. Boca
Raton, FL: CRC Press LLC.
Sambrook, J., Fritsch, E., & Maniatis, T. (1989). Molecular cloning: A laboratory manual, Second
Edition. New York, NY: Cold Spring Harbor Laboratory Press.
Sciarretta, M. J., Berlin, R. M., Smith, S. N., & Dawson Cruz, T. (2011). Evaluation of MCLAB
Buffer and NanoPOP-4™ as alternative consumables for Applied Biosystems capillary
electrophoresis. American Academy of Forensic Sciences Proceedings. 17, p. 42. Denver,
CO: Publication Printers.
Takara Bio Inc. (2013). SpeedSTAR™ HS DNA Polymerase: Product Manual, v201311Da. Shiga,
Japan: Takara Bio Inc.
Takara Bio Inc. (2014, November). SpeedSTAR DNA Polymerase. Retrieved from ClonTech:
http://www.clontech.com/takara/US/Products/PCR_Products/High_Speed_PCR/SpeedS
TAR_HS_DNA_Polymerase?sitex=10031:22372:US
Thermo Fisher Scientific. (2013, October). Thermo Scientific Piko Thermal Cycler: User Manual,
Rev. 1.3. Vantaa, Finland: Thermo Fisher Scientific.
University of Rhode Island. (2015, April 12). Eppendorf Mastercycler ep Gradient S. Retrieved
from University of Rhode Island: http://web.uri.edu/gsc/eppendorf-mastercycler-ep-
gradient-s/
Vallone, P. M., Hill, C. R., & Butler, J. M. (2008, December). Demonstration of rapid multiplex
PCR amplification involving 16 genetic loci. Forensic Sci Int Genet, 3(1), 42-45.
Verheij, S., Harteveld, J., & Sijen, T. (2012, March). A protocol for direct and rapid multiplex PCR
amplification on forensically relevant samples. Forensic Sci Int Genet, 6(2), 167-176.
238
Walsh, D. J., Corey, A. C., Cotton, R. W., Forman, L., Herrin, J. G., Word, C. J., & Garner, D. D.
(1992, March). Isolation of deoxyribonucleic acid (DNA) from saliva and forensic science
samples containing saliva. Journal of Forensic Sciences, 37(2), 387-395.
Walsh, P. S., Erlich, H. A., & Hilguchi, R. (1992, May). Preferential PCR amplification of alleles:
Mechanisms and solutions. Genome Research, 241-280.
Wang, D. Y., Chang, C.-W., & Hennessy, L. K. (2009). Rapid STR analysis of single source DNA
samples in 2 h. Forensic Sci Int Genet Supplemental Series, 2(1), 115-116.
Wang, Y., Jingyue, J., Carpenter, B. A., Atherton, J. M., Sensabaugh, G. F., & Mathies, R. A.
(1995, April). Rapid sizing of short tandem repeat alleles using capillary array
electrophoresis and energy-transfer fluorescent primers. Analytical Chemistry, 67(7),
1197-1203.
Wang, Z., Luo, H., Pan, X., Liao, M., & Hou, Y. (2012, May). A model for data analysis of
microRNA expression in forensic body fluid identification. Forensic Sci Int Genet, 6(3),
419-423.
Wasserstrom, A., Frumkin, D., Davidson, A., Shpitzen, M., Herman, Y., & Gafny, R. (2013,
January). Demonstration of DSI-semen--A novel DNA methylation-based forensic semen
identification assay. Forensic Sci Int Genet, 7(1), 136-142.
Westring, C. G., Kristinsson, R., Gilbert, D. M., & Danielson, P. B. (2007, September). Validation
of reduced-scale reactions for the Quantifiler™ Human DNA Kit. Journal of Forensic
Science, 52(5), 1035-1043.
Wisconsin State Crime Laboratory. (2013). Comparative DNA recovery of cotton-tipped
applicators with the DNA IQ™ extraction procedure. 12th Annual Advanced DNA
Technical Workshop-East. Amelia Island, FL.
Witt, S., Neumann, J., Zierdt, H., Gébel, G., & Röscheisen, C. (2012, February). Establishing a
novel automated magnetic bead-based method for the extraction of DNA from a variety
of forensic samples. Forensic Sci Int Genet, 6(5), 539-547.
Yang, H.-C., Liang, Y.-J., Huang, M.-C., Li, L.-H., Lin, C.-H., Wu, J.-Y., Chen, Y.-T., Fann, C. S. (2006,
August). A genome-wide study of preferential amplification/hybridization in microarray-
based pooled DNA experiments. Nucleic Acids Research, 34(15), e106.
Yang, H.-C., Pan, C.-C., Lu, R. C., & Fann, C. S. (2005, January). New adjustment factors and
sample size calculation in a DNA-pooling experiment with preferential amplification.
Genetics, 169, 399-410.