High Throughput Workflow for Global Metabolomic Profiling ... · High Throughput Workflow for Global Metabolomic Profiling with RapidFire online SPE/MS system ... SPE/MS 6/23/2015

High Throughput Workflow for Global Metabolomic Profiling with RapidFire online

SPE/MS system

Michelle Romm, Ph.D.Applications Scientist,

Agilent Technologies, Inc.

6/23/2015

1

What is High Throughput?

6/23/2015

2

# of Samples

N=1

N=100

N=1,000

N=10,000

LC/MS

30 mins

50 hrs

20.8 days

208.3 days

RapidFire/MS

14 secs

23 mins

3.8 hrs

38.8 hrs (1.6 days)

120X Faster

Why High Throughput Profiling?

6/23/2015

3

Method Development Fast Turnaround Large Population StudiespH

Solvents/AdditivesChemistry

Timings/Flows

N= 5505005000

QC/QAClass Prediction

Is Faster Better?When

6/23/2015

4

RapidFire/MS System

RapidFire 365 Online SPE/MS System

For Research Use Only. Not for Use in Diagnostic Procedures.

6/23/2015

5

BenchBot plate handling robot63 plate capacity (384 well plates = 24,192

samples)Off-deck reach capability

Quaternary pumpsSolvent switching capability Capable of running multiple disparate methods in

one batch

12 position automated cartridge changer Automated method development

Perform multiple injections of a single sample with different SPE cartridges and mobile phasesC18, C8, C4, CN, Phenyl, HILIC, Graphitic

Carbon Simply run multiple samples overnight and

select your ideal conditions the next morning

Aspirate Load/Wash Elute Re-equilibrate

600ms 3000ms 8000ms 1000ms

Method Development: Refining the Experiments

• Cartridge chemistries

• Ionization mode

• Loading solvents

• Elution solvents

One solution doesn’t fit everyone, this system allows FLEXIBILITY!

6/23/2015

6

Comparison of Chemistries: C18 vs. HILIC SPE/MS (+ESI) in Urine

526 metabolites annotated by C18 SPE/MS

270 metabolites annotated by HILIC SPE/MS

49 annotated metabolites common between HILIC and C18SPE/MS

6/23/2015

7

HILIC221

C18477

49

Utilize multiple unique chemistries to increase the metabolite coverage

747 unique metabolites (100% reproducibility; CV<15%)

Comparison of ESI Mode: a Small Polar Library

6/23/2015

8

Red- NEGBlue- POS

• ~100 pure endogeneous human metabolite standards• small organic acids, biogenic amines, nucleotides, amino acids

• Custom packed HILIC material

Comparison of Loading Solvents: HILIC Normal Phase

6/23/2015

9

ACN:Water (95:5)

Hexane:IPA (90:10)

Methanol

Lactic Acid, -ESI

Comparison of Elution pH

6/23/2015

10

MeOH:IPA (50:50) +0.1% Formic Acid

MeOH:IPA (50:50) +10mM Ammonium Acetate pH 9.0Lactic Acid, -ESI

High Throughput Metabolomics Profiling WorkflowBatch Data Processing Data Analysis

Verification/ Confirmation/

Validation

High ThroughputQuantitation

High ThroughputAcquisition

MassHunter Profinder

(NEW)

.d data input

RecursiveMFE

Features.cef output

Mass ProfilerProfessional

(MPP)

Statistics

Database Search

Pathway Analysis

Potential Targets

Pooled Sample/Random Samples

LC/MSAcquisitionFindbyFormulaVerify Targets

LC/MS/MSAcquisition

Metlin Library/SimLipid/

MSC

Confirm Targets

Validate withReference Standard

RapidFire 365

iFunnel 6550 Q-TOF

RapidFire 365

iFunnel 6490 QQQ

6/23/2015

11

Human Urine Metabolomics Profiling: Sample Prep Human urine, collected and pooled from healthy subjects (males and females), was purchased

from Innovative Research Inc. (Novi, MI)

Filter 400µL urine through a 0.22µm Ultrafree-MC® centrifugal filter units (Millipore) bycentrifugation @ 12,000xg for 20min @ 4°C

Pipette 50µL of filtered urine (10 replicates) into the Amicon Ultra-0.5mL Centrifugal Filter (3kDaCut-off Ultracel-3 membrane) from Millipore

Dilute with 350µL of Milli-Q quality water containing 0.2% Acetic acid, 1mM Ammonium acetate,and centrifuge @ 12000xg, 4°C for 30min to concentrate

The flow-through represents the metabolomics sample

Inject 10µL of flow-through for SPE/MS as well as LC/MS/MS analysis

Human urine equivalent to 1.25µL was used for the metabolomics profiling studies

6/23/2015

12

TIC for 10 Human Urine Samples (C18 SPE/MS)

6/23/2015

13

Single Injection of Urine Sampleon C18 RapidFire SPE/MS

6/23/2015

14

6x10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

1.9

2

2.1

2.2

2.3

2.4

Counts vs. Acquisition Time (min)

0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23

Male Urine+ESI

A Single Feature in Urine on C18 RapidFire SPE/MS

6/23/2015

15

4x10

0

0.25

0.5

0.75

1

1.25

1.5

1.75

2

2.25

2.5

2.75384.0461 (M+H)+

385.0466 (M+H)+

386.0505 (M+H)+406.0273 (M+Na)+

Counts vs. Mass-to-Charge (m/z)

379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412

Male Urine

5x10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

0.147 Cpd 323: 0.147


0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.2 0.21 0.22 0.23 0.24 0.25 0.26 0.27 0.28 0.29

EIC (C18, Pos ESI)Male Urine

ProFinder for High Throughput Batch Analysis of Samples

6/23/2015

16

Mass Profiler Professional (MPP)

6/23/2015

17

Cluster Analysis Filter by all 10 reps, abundance >2, CV < 15%

C18-PosESI

Annotation of Creatinine in Urine Samples by C18 RapidFire SPE/MS

6/23/2015

18

Select High Abundant Metabolites in Human Urine Annotated by SPE/MS

6/23/2015

19

Compound RF-C18 MaleHippuric acid 5158031Pirimicarb 3255466Enprofylline 496924Lys Asn Asp 4201013-Hydroxyhippuric acid 410564Pinacidil 359491Nadolol 3572284-(1-Hydroxy-2-(methylamino)ethyl)phenol 299538Leu Gln Lys 287646Triciribine 2859323-Amino-2-naphthoic acid 277964L-4-Hydroxy-3-methoxy-a-methylphenylalanine 2660331-Methylxanthine 262927AFMK 229545Creatinine 2252744-Chloroacetophenone 217597Jasmolone glucoside 209979Istamycin C1 192524N-Acryloylglycine 183487N2,N2-Dimethylguanosine 175089N(alpha)-t-Butoxycarbonyl-L-leucine 171812beta-Snyderol 1351772-Methylguanosine 126988

6/23/2015

20

Metlin Library Check for Isomers/Isobars

LC/MS/MS Confirmation of Hippuric acid

MFG for MS/MS Spectral PeaksHippuric acid

CID@10

CID@20

CID@40

Hippuric acidMS/MS @ 40Metlin Spectral Library

RapidFire C18 SPE/MS +ESI

6/23/2015

21

Male Urine

6/23/2015

22

300 newborn dried blood spot cards (males and females), were obtained fromthe state of California

3.2mm punch was taken from center of each spot and extracted in 200uL of icecold 80:20 Methanol:Water containing internal standards followed bycentrifugation @ 12,000xg for 10min @ 4°C

Supernatant was passed through a Phenomenex Phree SPE plate to removephospholipids

200uL of 20:80 Methanol:Water was added to flow through

10uL of final solution (equivalent to ~1.5uL of blood) was analyzed on the AgilentRapidFire in-line SPE system using Imtakt Scherzo SM-C18 mixed mode andbare silica HILIC cartridges in positive and negative ionization modes

Dried Blood Spot Metabolite Profiling: Sample Prep

6/23/2015

23

12

14

16

18

20

22

2413C Glutamine%CV = 7.6

13C Glucose%CV = 8.2

300 blood spots

Sig

nal I

nten

sity

What is the variability? Internal Standards

6/23/2015

24

10

12

14

16

18

20

22

Glutamate

10

11

12

13

14

15

16

17

Methionine

10

11

12

13

14

15

16

17

18

19

Histidine

10

12

14

16

18

20

22

24

Free Carnitine

10

11

12

13

14

15

16

17

18

19

20

Propionylcarnitine

10

11

12

13

14

15

16

17

18

Hexanoylcarnitine

Metabolite Verification

6/23/2015

25

PCA Analysis of DBS: Can we see differences?

Present in 75%, not present in solvent or blank blood card, abundance >10,000, fold change >2

SMC18-PosESI

6/23/2015

26

Gala Fuji Gold Delicious

Granny Smith

Red Delicious

75 apples total, 5 apples of each variety from 3 different stores

Peel was carefully collected from each apple with care to avoid removing any of theapple flesh or harvest from damaged areas

Peels were cryomilled, lyophilized and then extracted in methanol at -20C for 12 hours

Inject 10µL of sample for SPE/MS

Class Prediction Analysis: Sample Prep

6/23/2015

27

6x10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

Cpd 1: 0.128: +ESI ECC Scan Frag=175.0V Inj00008-Apple Peel-C18-A3

0.128 Cpd 1: 0.128


0.11 0.12 0.13 0.14 0.15 0.16

7x10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

1.6

1.7

1.8

+ESI TIC Scan Frag=175.0V Inj00008-Apple Peel-C18-A3-F-9.d


0.11 0.12 0.13 0.14 0.15 0.16

6x10

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

Cpd 1: 0.121: +ESI ECC Scan Frag=175.0V Inj00010-Apple Peel-GC-A5

0.121 Cpd 1: 0.121


0.11 0.12 0.13 0.14 0.15

7x10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

+ESI TIC Scan Frag=175.0V Inj00010-Apple Peel-GC-A5-GS-13.d


0.11 0.12 0.13 0.14 0.15

6x10

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

Cpd 1: 0.120: +ESI ECC Scan Frag=175.0V Inj00011-Apple Peel-E9-RD-

0.120 Cpd 1: 0.120


0.11 0.12 0.13 0.14 0.15 0.16

7x10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

+ESI TIC Scan Frag=175.0V Inj00011-Apple Peel-E9-RD-10.d


0.11 0.12 0.13 0.14 0.15 0.16

C18 Fuji

Graphitic CarbonGranny Smith

HILICRed Delicious

TIC 1

ECC 1

TIC 2 TIC 3

ECC 2 ECC 3

Different Cartridge Profiles on RapidFire SPE/MS

6/23/2015

28

ProFinder: Batch Analysis

6/23/2015

29

Features Detected

Variety C18Graphitic Carbon HILIC

Fuji 1017 404 229

Gala 991 411 227

Golden Delicious 987 411 227

Red Delicious 1005 421 229

Granny Smith 996 419 228

Data Analysis Workflow: Summary

Frequency 100% in one group, not present in blank, One-Way ANOVA (p<0.05, Benjamini-Hochberg corr)

Which features distinguish between the apple varieties?

6/23/2015

30

Partial Least Squares Discriminate Analysis Model for C18

R2 = 0.85Q2 = 0.57

Fuji

Gala

Gold D

Granny S

Red D

6/23/2015

31

R2 = 0.75Q2 = 0.35

Fuji

Gala

Gold D

Granny S

Red D

Partial Least Squares Discriminate Analysis Model for HILIC

6/23/2015

32

Model Building Workflow: Prediction Results

6/23/2015

33

RapidFire - Ion Mobility QTOF

System sensitivity optimized using electrodynamic ion funnels to focus and transmit ions

Ion Mobility resolution optimized while maintaining QTOF performance (mass resolution and accuracy)

Ion Fragmentation can be selected using standard QTOF collision cell (CID)

Bandwidth of QTOF data acquisition and processing channel was increased by 10 fold to accommodate the ion mobility data

IM-QTOF Instrument Overview

34

6/23/2015

35

Wash Blank

Trisaccharides

Melezitose

Raffinose

Why Ion Mobility: Separation!Resolving Structural Sugar Isomers C18H32O16

Tiabendazole

Imazaquin

Why RapidFire-IMQTOF: Sensitivity!

6/23/2015

37

Metabolites extracted from mice plasma

Metabolites extracted from human urine

Isomeric Separation: m/z=330.228

Same nominal mass: m/z=138.054, 138.129

Isomeric Separation: m/z=663.485

Analysis of Metabolomes from Different Matrices

Take Home Message…

6/23/2015

38

Using RapidFire/MS for metabolomics profiling provides a high throughput workflow that is flexible, efficient,

reproducible and produces high fidelity data

Agilent Technologies• Dan Cuthbertson• Sumit Shah

UC San Diego• Jeramie Watrous• Mohit Jain

PNNL• Xing Zhang• Erin Baker

6/23/2015

39

Acknowledgements

Documents

High Throughput Workflow for Global Metabolomic Profiling ... · High Throughput Workflow for Global Metabolomic Profiling with RapidFire online SPE/MS system ... SPE/MS 6/23/2015