41
Bioinformatics and Computational Biology Humberto Ortiz Zuazaga University of Puerto Rico High Performance Computing facility July 16, 2009

Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Bioinformatics and Computational Biology

Humberto Ortiz Zuazaga

University of Puerto RicoHigh Performance Computing facility

July 16, 2009

Page 2: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Bioinformatics

“The creation and advancement of algorithms, computational andstatistical techniques, and theory to solve formal and practicalproblems posed by or inspired from the management and analysisof biological data.” — Wikipedia

Page 3: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Computational biology

The application of computers to the collection, analysis, andpresentation of biological information.

Page 4: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Electrophysiological data collection

Proc. Natd Acad. Sci. USA 78 (1981) 7807

14 16 181 a

Temperature, 0C

805 12 14 16 18 20 22 24 26a a I I A a a

Temperature, 0C

B C

.....

*;,,, ~ ~~~~~~~~~~~~~~~........ ..........

FIG. 1. Sodium bisulfite produces an increase in mepc decay time.(A) Each point is an average of records from five cells (130 mepc recordsper cell). Records were taken at the three temperatures with a mem-

brane holding potential of 100 mV. Inward current was plotted as an

upward deflection. o, Control values; e, bisulfite-treated endplates.(B and C) Sample records from which A was constructed. Each figureis a computer-generated average from 130 mepcs taken from one end-plate. (Inset) mepc decay (80-20%) as a semilogarithmic plot. All am-plitudes were normalized. Decay time constants were 1.93 msec forcontrol (B) and 2.59 msec for bisulfite-treated (C) endplates. Temper-ature, 150C.

then the reagents were applied in Ringer's solution containingmethanesuilfonyl fluoride. The mepc decay, already prolongedby the methanesulfonyl fluoride was further prolonged by thereagents (Fig. 3). As a further test, experiments were conductedin which iontophoretic pulses ofAcCho and carbachol were ap-plied before and after the application of bisulfite or diamide.If the action of the reagents were on the esterase, one wouldexpect to see an increase in the decay time of the response toAcCho but not to carbachol because the latter is not hydrolyzedby the esterase (21). When short (1-2 msec) pulses of AcChoand carbachol were applied iontophoretically to the endplate,there was an increase in the decay time of both responses.

Therefore, an action of the reagents on acetylcholinesterasedoes not seem likely.An alternative explanation for the increase in mepc decay

A a'_^ ., d

:*

-

..,

......

B ;.A*: .§I, ..0 N

:-

FIG. 2. Diamide produces an increase in mepc decay time. Con-ditions were as in Fig. 1. mepc decay time constants for traces are 1.32msec for control (B) and 1.82 msec for diamide-treated (C) endplates.

time would be to assume that the reagents produce an increasein the open time of the AcCho-activated channels because therate-limiting step in the mepc decay is believed to be the re-

laxation kinetics of the AcCho receptor (ref. 22; however, see

refs. 23 and 24). When the single-channel parameters were

determined by spectral analysis ofAcCho noise, however, therewas no statistically significant change in either mean channelopen time or conductance after treatment of the endplate witheither bisulfite or diamide (Fig. 4 and Table 1). Although av-

erage ;pectra (Table 1) and average Trmepc (Figs. 1 and 2) for con-trol endplates show some variability, such variability is not sig-nificant when compared to the difference in time constantsfound between control and reagent-treated fibers in the same

preparation.In addition to the increased decay time of the mepc, an in-

crease in the time to peak was also observed in treated end-plates. This effect was not quantitated but was sufficiently largeto be detected easily by eye. No change in the peak amplitudeof the mepc was found after treatment with the reagents. In thepresent series of experiments, an increase in the input resis-tance was observed in some fibers after treatment with bisulfiteor diamide. This effect has been observed for bisulfite (25). Theincrease in input resistance may partly account for the increasein the average mepp amplitude but cannot affect either the am-plitude or the time course of the mepc recorded under voltageclamp.

B

7-.

-, :1

9,

::.

FIG. 3. Increase in mepc decay is not due to de-struction of acetylcholinesterase. mepc after 1 hr inmethanesulfonyl fluoride (A) followed by a 10-mintreatment with bisulfite applied in methanesulfonylfluoride/Ringer's solution (B). Records were taken asfor the mepc averages in Figs. 1 and 2.

A

cql

CDCDC5

CD'"' Aco

00

12 20 22 24 26. . * .~~~~~~~~~~~~~

Neurobiology: Steinacker and Zuazaga

IS

K4)0

1.ll-.

.W .. At l anta.'

Steinacker A, Zuazaga DC. Changes in neuromuscular junctionendplate current time constants produced by sulfhydryl reagents.Proc Natl Acad Sci U S A. 1981 Dec;78(12):7806–7809.

Page 5: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Data collection system

Digital Equipment Corporation (DEC) PDP-11. Replaced highspeed camera pictures of oscilloscope followed by manualmeasurement of trace heights encoded on a deck of punched cardsfor processing by IBM mainframe in Facundo Bueso.

Page 6: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Electrophysiological simulation

MEMBRANE CURRENT IN NERVE

the sodium and potassium conductances to time and membrane potential.Before attempting this we shall consider briefly what types of physical systemare likely to be consistent with the observed changes in permeability.

Outside

T INT TK ItECM~~~~~~+ +

Inside )

Fig. 1. Electrical circuit representing membrane. RB =l/gNa; RK= l/9K; RI= 1/#1. RNw andRK vary with time and membrane potential; the other components are constant.

The nature of the permewablity change8At present the thickness and composition of the excitable membrane are

unknown. Our experiments are therefore unlikely to give any certain informa-tion about the nature of the molecular events underlying changes in perme-ability. The object of this section is to show that certain types of theory areexcluded by our experiments and that others are consistent with them.The first point which emerges is that the changes in permeability appear to

depend on membrane potential and not on membrane current. At a fixeddepolarization the sodium current follows a time course whose form is inde-pendent of the current through the membrane. If the sodium concentrationis such that ENaB<E, the sodium current is inward; if it is reduced untilENa > E the current changes in sign but still appears to follow the same timecourse. Further support for the view that membrane potential is the variablecontrolling permeability is provided by the observation that restoration of thenormal membrane potential causes the sodium or potassium conductance todecline to a low value at any stage of the response.The dependence of 9Na and g9 on membrane potential suggests that the

permeability changes arise from the effect of the electric field on the distribu-tion or orientation of molecules with a charge or dipole moment. By this wedo not mean to exclude chemical reactions, for the rate at which these occurmight depend on the position of a charged substrate or catalyst. All that isintended is that small changes in membrane potential Would be most unlikely

501

A. L. Hodgkin and A. F. Huxley. A quantitative description ofmembrane current and its application to conduction and excitationin nerve. J Physiol. 1952 August 28; 117(4): 500–544.

Page 7: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Electrophysiological verification

MEMBRANE CURRENT IN NERVE

RESULTS

Membrane action potentialsForm ofaction potential at 60 C. Three calculated miembrane action potentials,

with different strengths of stimulus, are shown in the upper part of Fig. 12.Only one, in which the initial displacement of membrane potential was 15 mV,is complete; in the other two the calculation was not carried beyond the middleof the falling phase because of the labour involved and because the solution

110100

80--- 70E 60::.501 40

30 1is3 4

20100

msec

80~70E 60>.501 40

20100

0 1 2 3 4 5 6msec

Fig. 12. Upper family: solutions of eqn. (26) for initial depolarizations of 90, 15, 7 and 6 mV(calculated for 60C). Lower family: tracings of membrane action potentials recorded at6' C from axon 17. The numbers attached to the curves give the shock strength inm1Lcoulomb/cm2. The vertical and horizontal scales are the same in both families (apart fromthe slight curvature indicated by the 110 mV calibration line). In this and all subsequentfigures depolarizations (or negative displacements of V) are plotted upwards.

had become almost identical with the 15 mV action potential, apart from thedisplacement in time. One solution for a stimulus just below thireshold is alsoshown.The lower half of Fig. 12 shows a corresponding series of experimental

membrane action potentials. It will be seen that the general agreement isgood, as regards amplitude, form and time-scale. The calculated actionpotentials do, however, differ from the experimental in the following respects:(1) The drop during the first 01 msec is smaller. (2) The peaks are sharper.

525

Computed action potentials on top, experimental action potentialson bottom. Awarded the 1963 Nobel Prize in Physiology orMedicine.

Page 8: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Moore’s law

Image from Wikimedia commons by Wgsimon, used withpermission.

Page 9: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Larger scale simulations

N. Sperelakis, H. Ortiz-Zuazaga, and J. B. Picone. Fastconduction in the electric field model for propagation in cardiacmuscle. Innov. et Tech. en Biol. et Med., 12(4):404-414, 1991.

Page 10: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Larger scale results

Page 11: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

The end of Moore’s law

Where’s my 4 GHz processor?

Page 12: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Simulation of groundwater contamination

A GRACE interface for GRASS. John Franco and HumbertoOrtiz-Zuazaga. U.S. Army Corps of Engineers, $75,000,1994–1995.

Page 13: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Neural network processing of cardiotocograms

B. E. Rosen, D. Soriano, T. Bylander, and H. Ortiz-Zuazaga.“Training Neural Networks to recognize Artifacts andDecelerations in Cardiotocograms.” AAAI Symposium on ArtificialIntelligence in Medicine. pp. 149–153, 1996.

Page 14: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Genetic Mapping

I Goal: The determination of orders and distances amongmarkers on a chromosome based on the observed patterns ofinheritance of the alleles of the markers in three generationpedigrees.

I Problem: For a variety of reasons the genotypic information isnot complete, and not all crosses in human pedigrees areinformative. In addition, the time required to order markersgrows exponentially with the number of markers.

I Solution: Only use “good” markers to make maps. Biologistsalready have a notion of a “framework” map, a map of asubset of the markers which has very high odds againstinversion of adjacent markers.

Page 15: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Meiotic breakpoints

From http://www.stanford.edu/group/Urchin/

Page 16: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

A genotyped pedigree

Page 17: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Counting obligate breaks as an estimate of genetic distance

A simple estimate of the genetic distance between two markers isthe number of observed recombinations between the markers in thedata set. For the first two markers in our sample pedigree wewould have:

UT851 U M U M U M U M U M U MUT1398 P P P M M M P P P M M M

Breaks 1 1

for a total of 2 breaks.This technique based on counting the number of recombinations isknown as meiotic breakpoint analysis (BPA).

Page 18: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Selecting genetic markers with wclique

I Each marker becomes a node of a graph.

I The weight of the node is the total count of P and M phasesfor this marker.

I Two nodes in the graph are connected by an edge whoseweight is the number of breaks between the correspondingmarkers.

Page 19: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

A small distance graph

Page 20: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Finding framework markers is a graph problem

I Finding a good set of framework markers is now a graphproblem: find a set of nodes with maximal weight where allthe nodes are connected by an edge of weight e or higher.

I This graph problem is called Maximal Weighted Clique(MWC).

Page 21: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

The maximal weighted clique problem is NP-complete

I The MWC is a well known graph problem, extensively studiedin computer science. Unfortunately, it belongs to the class ofNP-complete problems, for which there is unlikely to be anefficient algorithm.

I Building a linear map by ordering genetic markers so as tominimize the number of recombination events in a set ofgametes can also be cast as a graph problem, the travelingsalesman problem (TSP), which is also NP-complete.

Page 22: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

But I still need a map

I Exact algorithms can work on small sets of markers.

I Local search techniques can find near optimal solutions forsome of these problems, at the cost of not knowing if anoptimal solution was ever found. The best heuristics for TSPcan find a solution with 1.05 times the optimal cost.

I A change in the formulation of the problems can enable otheralgorithms to be used. For example, if the data had no errors,was complete, and no double recombination events occurred,ordering genetic markers would be equivalent to theconsecutive ones problem (C1P) for which there are lineartime algorithms.

Page 23: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Searls plot of unselected markers

Page 24: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Searls plot of wclique-selected markers

Page 25: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Comparison of MLA maps of hand-selected andwclique-selected markers

Page 26: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

References

1. H. Ortiz-Zuazaga, and R. Plaetke. Screening geneticmarkers with the maximum weighted clique method. Abstractpresented at Genome Mapping and Sequencing. Cold SpringHarbor, May 1997.

2. S.L. Naylor, R. Plaetke, H. Ortiz-Zuazaga, P. O’Connell, B.Reus, X. He, R. Linn, S. Wood, and R.J. Leach. Constructionof Framework and Radiation Hybrid Maps of Chromosomes 3and 8. Abstract presented at Genome Mapping andSequencing. Cold Spring Harbor, NY, May 1997.

Page 27: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

“Moore’s law” for sequence data

100000

1e+06

1e+07

1e+08

1e+09

1e+10

1e+11

1e+12

1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012

Base

pai

rs

Year

Growth of the Genebank sequence database

From the June 15 2009 NCBI-GenBank Flat File Release 172.0

Page 28: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Gene expression networks

I Complete genomes available for several species.

I 40,000 human genes, many already sequenced.

I microarrays can measure expression levels for ALL GENES ina single assay.

Page 29: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Microarray image

Reproduced from www.molecularstation.com

Page 30: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Microarray data

Raw log ratio vs log intensity for two color microarrays.

Page 31: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Microarray analysis

●●

●●

●●

●●

● ●

●● ●

●●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●

●●●●

●●

●●

● ●●

●●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●●●

● ●●

●●

● ●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●●

●●

●●

●● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

● ●●

●●

●●●●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●● ●

● ●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●●●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●

● ●●

● ●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●●●

●●

●●●

●●●

●●

● ●

●●

●●

●●

●●●●●

●●

●●

● ●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

● ●●

●●●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

●●

●● ●

●●

●●

●●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●●

● ●

●●

●●

●●

●●●

● ●

●●

●●●

●●

●●

●●

● ●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

● ●●

●●

●●

●●

●●

●●

●●●

●● ●●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●●●●

●●●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●●●●

●●

●●

● ●

●●

●●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●● ●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●●

●●

●●●

● ●

● ●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●● ●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●● ●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●●●

●●

● ●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●● ●●●

●●

●●●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●● ●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●●

●●

● ●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●●●

●●

● ●

●●●

●●

●●●

●●

●●●

● ●

●●●

●●

●●

●●

●● ●●

●●●●●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●●

●●●

●●

●●

●●

●● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

● ●●

●●

●●

●●

●●

●●

● ●

●●●

● ●●

●●

●●

●●●

●●

●●●

●●

● ●

●●

● ●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

● ●●

● ●

●●

●●

●●

●●

● ●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

● ●●

● ●●

●●

●●

● ●

●●

●●

● ●

●●

● ●

●● ●

● ●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●●●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●● ●●

● ●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●● ●

● ●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

● ●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●●

●●

●●

●●

●●●●

●●

●●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

● ●

●●●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●● ●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

● ●●

●●●●●

● ●

●●

●●

●●

●●

●●●● ●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●● ● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●●

●●

●●●

●●

● ●

●●

●●

●●

●●●●●

●●

●●●●

●●●

●●

●●

●●

●●

● ●

● ●

●●●

●●

●●

●●

● ●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

● ●

●●●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●●

●●●

● ●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●● ●

●●

●●

●●

●●

● ●●

●●

●●

●●●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●●

●●●

●●

●●

● ●

●●

●●● ●

●●●

●●●

●●

●●

●●

●●●

●●

●● ●

● ●

●●

●●●

●●

●●

● ●

●●

● ●●

●●

●●

●●

●●●●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●●

●●

●●

● ●

●●

● ●

●● ●

●●

●●

● ●

●●●●

●●

●●

● ●

●●

● ●

●●

●●

●●●

●●

● ●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

● ●

● ●

●●●

●●●

● ●

●●

●●

●●

●●

● ●●

●●

● ●

●●

●●

●●

●●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●●

● ●

●●

●●

●●

● ●

●●

●●

● ●

●●●

●●●

● ●

●● ●

●●

●●

●●

● ●

●●

●●

●●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●●

●●

● ●

● ●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●●

●●●

●●

●●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●●

●●

●●

●●

●●

●●

●●

●●●●

●●●

●●

●●

●●

●●

●●●

● ●●

●●

●●

●●

● ●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●●

●●

●●

● ●

●●

●●

●●

●●●

● ●

●● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●●

●●

●●●

●●

●●

●●

●● ●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

● ●

●●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●●

●●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●●

● ●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●

●●

● ●●

●●

●●

●●

●●

●●●

● ●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

● ●

●●

●●

●●

●●●

●●

●●●

● ●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●

●●

●●

●●

● ●

● ●●

● ●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

● ●

●●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

● ●●

●●

●●

●●

●●

●●

● ●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●●

●●

●●

●●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●

●●

●●

●● ●●●

● ●

●●

● ●

●●●

●●

●●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●●

● ●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●●

●●

●●

●●

●●●●

● ●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●●

●● ●

●●●

●●

●●●

●● ●

●●●

●●

● ●

●●● ●

●●

●●

●●

●●

●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

● ●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

● ●

●● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●● ●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●●●

●●

●●●

●●

●●●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●●●

●●

●●

●●

●●

●●●●

●●●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●●

●●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●

●●

●●●

● ●

●●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●● ●

● ●

●●

● ●

●●●

●●

●●

●●

● ●

●●●

●●

●●

●●

●●

●● ●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●●●

●●

●●●

●●

●●

●●

●●

●●

●●● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●●

●●

●●●

●●

●●●

●●

● ●

●●

●●

●●●

●●

●●

●●

●●

● ●

●●

●●

●●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●

●●

●●

●●

● ●

●●

●●

● ●●

● ●●

●●

●●

●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●● ●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●●

●●

●●●

●●

●●

●●●

●●●

●●

●●

●●

● ●

●●

●●

●●●

●●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

● ●

● ●

●●

●●

●●●

●●

● ●●

●●●

●●

●●

●●

●●●

●●

●●

● ●●●

●●

● ●

●●

●●

●●●

●●

●●

●●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

● ●

● ●

●●

● ●

●●

●●

●●

●●

●●

●●●

● ●

● ●

●●

●●

● ●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●●

●●●

●●

● ●

●●

●●

●●

● ●

●●● ●●

●●

● ●

●●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

● ●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●●●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●

● ●

●●

●●●

● ●●

●●

● ●

●●

●●

●●●●

● ●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

●● ● ●

●●

●●

●●

●●

●●

●●

●●

●●

●●

●●●

●●

●●

● ●

●●

●●

●●

●●

●●●

●●

●●

●●

●●

●●

●●

●●

−2 −1 0 1 2 3 4

−5

05

1015

Log Fold Change

Log

Odd

s

Find the differentially expressed genes.

Page 32: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

“Moore’s law” for microarrays

100

1000

10000

100000

1e+06

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Feat

ures

Year

Feature count on commercially available arrays

Page 33: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Boolean Genetic Network Model

We define Boolean Genetic Network Model (BGNM):

I A Boolean variable takes the values 0, 1.

I A Boolean function is a function of Boolean variables, usingthe operations ∧, ∨, ¬.

A Boolean genetic network model (BGNM) is:

I An n-tuple of Boolean variables (x1, . . . , xn) associated withthe genes

I An n-tuple of Boolean control functions (f1, . . . , fn),describing how the genes are regulated

Page 34: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Boolean genetic networks

2 4

1 3

f1 = 1f2 = 1f3 = x1 ∧ x2

f4 = x2 ∧ ¬x3

Page 35: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Previous results on Boolean networks

I Determining if a given assignment to all the variables isconsistent with a given gene network was shown to beNP-complete in [1] (by reduction from 3-SAT).

I In the worst case, 2(n−1)/2 experiments are needed

I If the indegree of each node (the genes that affect our targetgene) is bound by a constant D, the cost is O(n2D).

I For low D, [2] and [3] provide effective procedures for reverseengineering, assuming any gene may be set to any value.

Page 36: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Reverse engineering Boolean networks

1. Akutsu, S. Kuahara, T. Maruyama, O. Miyano, S. 1998.Identification of gene regulatory networks by strategic genedisruptions and gene overexpressions. Proceedings of the 9thACM-SIAM Symposium on Discrete Algorithms (SODA 98),H. Karloff, ed. ACM Press.

2. Ideker, T.E., Thorsson, V., and Karp, R.M. 2000. Discoveryof regulatory interactions through perturbation: inference andexperimental design. Pacific Symposium on Biocomputing5:302-313.

3. S. Liang, S. Fuhrman and R. Somogyi. 1998. REVEAL, AGeneral Reverse Engineering Algorithm for Inference ofGenetic Network Architectures. Pacific Symposium onBiocomputing 3:18-29.

Page 37: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

The world’s smallest finite field

The integers 0 and 1, with integer addition and multiplicationmodulo 2 form the finite field Z2 = {{0, 1}, +, ·}.The operators + and · are defined as follows:

+ 0 1

0 0 11 1 0

· 0 1

0 0 01 0 1

Page 38: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Finite field equivalents to the Boolean operators

We can realize any Boolean function as an expression over Z2:

X ∧ Y = X · YX ∨ Y = X + Y + X · Y¬X = 1 + X

Page 39: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Finite field genetic networks

Any BGNM can be converted into an equivalent model over Z2 byrealizing the Boolean functions as sums-of-products andproducts-of-sums, then converting the Booleans to Z2. We nowhave a finite field genetic network (FFGN):

I An n-tuple of variables over Z2, (x1, . . . , xn) associated withthe genes

I An n-tuple of functions over Z2, (f1, . . . , fn), describing howthe genes are regulated

Page 40: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Publications

1. Ortiz-Zuazaga, H., Avino-Diaz, M. A., Laubenbacher, R.,Moreno O. Finite fields are better Booleans. Refereedabstract, poster to be presented at the Seventh AnnualConference on Computational Molecular Biology (RECOMB2003), April 10–13, 2003, Germany.

2. Humberto Ortiz-Zuazaga, Sandra Pena de Ortiz, OscarMoreno de Ayala. Error Correction and Clustering GeneExpression Data Using Majority Logic Decoding. Proceedingsof The 2007 International Conference on Bioinformatics andComputational Biology (BIOCOMP’07), Las Vegas, Nevada,June 25–28, 2007.

3. Humberto Ortiz Zuazaga, Tim Tully, Oscar Moreno.Majority logic decoding for probe-level microarray data.Proceedings of BIOCOMP’08 — The 2008 InternationalConference on Bioinformatics and Computational Biology, LasVegas, Nevada, July 13–17, 2008.

Page 41: Bioinformatics and Computational Biologyhumberto/presentations/hoz... · 2009-09-19 · Bioinformatics \The creation and advancement of algorithms, computational and statistical techniques,

Molecular phylogeny

Filipa Godoy-Vitorino, Ruth E. Ley, Zhan Gao, Zhiheng Pei,Humberto Ortiz-Zuazaga, Luis R. Pericchi, Maria A.Garcia-Amado, Fabian Michelangeli, Martin J. Blaser, Jeffrey I.Gordon, Maria G. Dominguez-Bello. Bacterial Community in theCrop of the Hoatzin, a Neotropical Folivorous Flying Bird. Appliedand Environmental Microbiology, October 2008, p. 5905–5912,Vol. 74, No. 19. doi:10.1128/AEM.00574-08