the basics - computational systems biology labcsbl.bmb.uga.edu › ~xyn › compbio-intro.pdf · an...

The BasicsThe Basicsccgtacgtacgtagagtgctagtctagtcgtagcgccgtagtcgatcgtgtgggtagtagctgatatgatgcgaggtaggggataggatagcaacagatgagcggatgctgagtgcagtggcatgcgatgtcgatgatagcggtaggtagacttcgcgcataaagctgcgcgagatgattgcaaagragttagatgagctgatgctagaggtcagtgactgatgatcgatgcatgcatggatgatgcagctgatcgatgtagatgcaataagtcgatgatcgatgatgatgctagatgatagctagatgtgatcgatggtaggtaggatggtaggtaaattgatagatgctagatcgtaggta…………………………………

genes

cell chromosome genome

protein metabolic pathway

Solving Biological Problems Solving Biological Problems Computationally Computationally

Ying Ying XuXuInstitute of Bioinformatics and Biochemistry and Institute of Bioinformatics and Biochemistry and

Molecular Biology DepartmentMolecular Biology DepartmentUniversity of GeorgiaUniversity of Georgia

http://http://csbl.bmb.uga.educsbl.bmb.uga.edu

Biology and ComputationBiology and Computation

biology computation

molecular biology is information science .......-- Leroy Hood, RECOMB’00

BioinformaticsBioinformatics(or computational biology)(or computational biology)

•• This interdisciplinary science This interdisciplinary science …… is aboutis aboutproviding computational support to providing computational support to studies studies on linking the behavior of cells, organisms on linking the behavior of cells, organisms and populations to the information and populations to the information encoded in the genomes.encoded in the genomes.– Temple Smith, Current Topics in Computational

Molecular Biology (2002)

ccgtacgtacgtagagtgctagtctagtcgtagcgccgtagtcgatcgtgtgggtagtagctgatatgatgcgaggtaggggataggatagcaacagatgagcggatgctgagtgcagtggcatgcgatgtcgatgatagcggtaggtagacttcgcgcataaagctgcgcgagatgattgcaaagragttagatgagctgatgctagaggtcagtgactgatgatcgatgcatgcatggatgatgcagctgatcgatgtagatgcaataagtcgatgatcgatgatgatgctagatgatagctagatgtgatcgatggtaggtaggatggtaggtaaattgatagatgctagatcgtaggta…………………………………

BioinformaticsBioinformatics

•• It is about using computational techniques toIt is about using computational techniques to–– interpret biological datainterpret biological data–– predict structures and functions of biological entities predict structures and functions of biological entities –– model the dynamic behavior of biological processes and systemsmodel the dynamic behavior of biological processes and systems–– ……....

•• People have used mathematical or computational People have used mathematical or computational techniques to solve biological problems since early 1900techniques to solve biological problems since early 1900’’ss–– e.g., evolution and genetic analyses by R.A. Fisher, J.B.S. e.g., evolution and genetic analyses by R.A. Fisher, J.B.S. HaldaneHaldane, ,

S. WrightS. Wright

•• So what is new?So what is new?

A Historical PerspectiveA Historical Perspective

•• Realization of the existence of Realization of the existence of ““genesgenes”” in our cells by in our cells by Hermann Hermann MMüüllerller, a student of Morgan, a student of Morgan’’s (1921)s (1921)

•• Understanding of the physical natures of genes byUnderstanding of the physical natures of genes by–– F. Sanger (e.g., 1949), F. Sanger (e.g., 1949), –– E. E. ChargraffChargraff (e.g., 1950), (e.g., 1950), –– J. J. KendrewKendrew (e.g., 1958)(e.g., 1958)

in 40in 40’’ and 50and 50’’ss


•• Understanding of the double helical structure of DNA by Understanding of the double helical structure of DNA by James Watson and Frances Crick in 1953James Watson and Frances Crick in 1953

•• Development of sequencing technology, first of proteins Development of sequencing technology, first of proteins and then of genomic, based on the work of and then of genomic, based on the work of –– F. Sanger on sequencing of insulin (1956), F. Sanger on sequencing of insulin (1956), –– W Gilbert and A. W Gilbert and A. MaxamMaxam on sequencing of Lactose operator on sequencing of Lactose operator

(1977)(1977)which demonstrated that the genetic sequence of a genome, including human’s, is sequence-able!


•• Development of the Development of the sciencescience of analyzing protein and of analyzing protein and DNA sequences, particularly inDNA sequences, particularly in–– protein sequence analyses and evolution byprotein sequence analyses and evolution by Margaret Margaret DayhoffDayhoff

(60(60’’s)s)–– phylogeneticphylogenetic analyses analyses andand comparative sequence analyses by comparative sequence analyses by W. W.

Fitch and E. Fitch and E. MargoliashMargoliash (1967)(1967) and by and by R. Doolittle (1983)R. Doolittle (1983)


•• Development of sequence comparison algorithmsDevelopment of sequence comparison algorithms–– Needleman and Needleman and WunschWunsch (1970)(1970)–– Smith and Waterman (1981)Smith and Waterman (1981)

•• Organization of biological data into databasesOrganization of biological data into databases–– GENBANK (1982) of DNA sequencesGENBANK (1982) of DNA sequences–– Protein Data Bank (PDB, 1973) of protein structuresProtein Data Bank (PDB, 1973) of protein structures

•• Computational methods for gene finding in genomic Computational methods for gene finding in genomic sequencessequences–– Work by Work by BorodovskyBorodovsky, , ClaverieClaverie, , UberbacherUberbacher from midfrom mid--8080’’s to s to

early 90early 90’’ss

A Historical PerspectiveA Historical Perspective•• Sequencing of Human and other genomesSequencing of Human and other genomes

•• Development of Development of ““highhigh--throughputthroughput”” measurement measurement technologies technologies –– microarraymicroarray chips for functional states of geneschips for functional states of genes–– twotwo--hybrid systems for proteinhybrid systems for protein--protein interactionsprotein interactions–– structural genomics for structure determinationstructural genomics for structure determination–– …………....

ccgtacgtacgtagagtgctagtctagtcgtagcgccgtagtcgatcgtgtgggtagtagctgatatgatgcgaggtaggggataggatagcaacagatgagcggatgctgagtgcagtggcatgcgatgtgatagctagatgtgatcgatggtaggtaggatggtaggt genes

A Historical PerspectiveA Historical Perspective•• These These ““highhigh--throughputthroughput”” probing technologies and probing technologies and

others are being used to generate enormous amounts of others are being used to generate enormous amounts of data data about the existence, the structures, the functional about the existence, the structures, the functional states, the relationships of biological molecules and states, the relationships of biological molecules and machineriesmachineries but but ……....

•• …….. what are these data telling us? .. what are these data telling us? ……....


•• So what is new?So what is new?

It is the amount & the type of biological data about the cellular states, and molecular structures and functions, generated by high-throughput technologies, that have driven the rapid advancement of bioinformatics!

An Example of Computation for BiologyAn Example of Computation for Biology

• Goal: to understand the phosphorus assimilation process in Synechococcus (a microbe in ocean) at molecular level– Observation: When the living environment is low in phosphate,

Synechococcus cells will turn on different machineries with higher affinity for in-taking particular forms of phosphate

– Question: how is this regulatory mechanism implemented at the molecular level?

Sensing device

Regulatory device

Action

1

Action

2

Action

3

Action

N

cell membrane

• Hypothesis driven research– Need to come up with a hypothesized model– Search for genes that may be involved in the

proposed model• perturbation (gene knock-out/mutation) and

phenotypic observations (data analysis)

– ……………– Model revision based on collected data, and

iterate!


• X years ago, …. to search for which genes are possibly involved in this process, researchers had to

– remove various parts of DNA sequence,– then observe if they may have any relevance

acggtcgtacgtacgtgttagccgataatccagtgtgagatacacatcatcgaaacacatgaggcgtgcgatagatgatcc.....

?

x

?

x

?

x

?

x

This could be a very lengthy process ……


• Since the inception of Human genome project (1986), computational scientists have developed computer programs to locate genes in a long stretch of DNA sequence– GRAIL, Gene-Scan, Glimmer, …….

• With gene-prediction programs, researchers only needed to knock-out regions predicted to be genes in their search for understanding of phosphorus assimilation process

acggtcgtacgtacgtgttagccgataatccagtgtgagatacacatcatcgaaacacatgaggcgtgcgatagatgatcc.....

genes

Great saving in time ……..


• Over the years, many genes have been thoroughly studied in different organisms, e.g., human, mouse, fly, …., rice, …

– their biological functions have been identified and documented

• Computational scientists have developed computer programs to associate newly identified genes to genes with known functions!– Existing methods can associate > 60% of newly identified

genes to genes with known functions

• Now, researchers only need to knock-out genes with possibly relevant functions in their search for understanding of a particular biological process …….

Saving even more time ……..


•• Deciphering a biological process, Deciphering a biological process, e.g., phosphorus assimilatione.g., phosphorus assimilation, , involves more than just gene finding/functional assignmentsinvolves more than just gene finding/functional assignments– interactions between gene products (proteins)– dynamic behavior of genes throughout a biological process– mechanistic issues re. how a protein implements its function – ………

•• Computational programs are springing out that can Computational programs are springing out that can – predict if two proteins are interacting with each other– model the dynamic behavior of metabolic pathways ….– predict protein structures to help answer mechanistic issues of

its function– …….. … and many more yet to come …..

Computation for BiologyComputation for Biologycomputation can help quickly

narrow down the search space

Searching a needle in a haystack …

Examples of Examples of ““Computation for BiologyComputation for Biology””

•• Suggesting functions of newly identified genesSuggesting functions of newly identified genes–– It was known that mutations of NF1 are associated with inheritedIt was known that mutations of NF1 are associated with inherited

disease neurofibromatosis 1; but little is known about the disease neurofibromatosis 1; but little is known about the molecular basis of the diseasemolecular basis of the disease

–– Sequence search found that NF1 is homologous to a yeast protein Sequence search found that NF1 is homologous to a yeast protein called called IraIra, which is a GAP, which is a GAP--type protein and known to regulate the type protein and known to regulate the function of a second type of protein called function of a second type of protein called RasRas

–– HypothesisHypothesis: NF1 regulates : NF1 regulates RasRas in human cell; followin human cell; follow--up up experiments verified this.experiments verified this.

NF1

Ras Ras

Ira


• Computer-assisted drug design• 3D structure models of G protein-coupled receptors were used to

computationally screen 100,000+ compounds as possible drug targets and 100 were selected

• Follow-up experiments confirmed a high hit rate of 12%-21%

OM Becker, et al, PNAS, 2004, 101:11304-11309


•• Computational analysis of Computational analysis of Plasmodium Plasmodium falciparumfalciparum metabolismmetabolism–– Plasmodium causes human malariaPlasmodium causes human malaria

–– computational prediction of metabolic computational prediction of metabolic pathways of plasmodiumpathways of plasmodium

–– computational simulations have helped to computational simulations have helped to identify 216 identify 216 ““chokepointschokepoints”” in this pathway in this pathway modelmodel

–– among all 24 previously suggested drug among all 24 previously suggested drug targets, 21 target at the targets, 21 target at the ““chokepointschokepoints””

–– among the three popular drugs for malaria, among the three popular drugs for malaria, they all targeted at the they all targeted at the ““chokepointschokepoints””

Yeh I. et. al. Genome Research 2004, 14:917-24

Computation for BiologyComputation for Biology

• Computation can help quickly reduce the search space of “searching a needle in a haystack”

• help rationally design experiments

• help to formulate new hypotheses

• …..

It is becoming an essential part of the tool kit of modern biology!

What Can Computation Do for BiologyWhat Can Computation Do for Biology

•• EvolutionEvolution--based computationbased computation ---- deriving structural, deriving structural, functional and evolutionary informationfunctional and evolutionary information–– Sequence alignment algorithmsSequence alignment algorithms for detection of for detection of homolgueshomolgues, functional , functional

inference, etc. inference, etc.

–– Multiple sequence alignment algorithmMultiple sequence alignment algorithm: detection of functional sites, : detection of functional sites, understanding of genome rearrangement, etcunderstanding of genome rearrangement, etc

–– Phylogeny reconstruction Phylogeny reconstruction algorithmalgorithm: estimation of evolutionary distance, : estimation of evolutionary distance, evolutionary tree reconstruction, etcevolutionary tree reconstruction, etc

–– Protein threading algorithmsProtein threading algorithms: identification of remote homologues, : identification of remote homologues, prediction of protein 3D structures, etc.prediction of protein 3D structures, etc.

–– ……....


Multiple protein sequence alignment

conserved sites and hence possibly functional sites

phylogenetic tree


•• PhysicsPhysics--based computationbased computation –– helping to derive helping to derive information about functional mechanismsinformation about functional mechanisms–– AbAb initioinitio folding of protein structuresfolding of protein structures –– calculation of structural calculation of structural

conformation that has the lowest potential energyconformation that has the lowest potential energy

–– Protein docking calculationProtein docking calculation –– docking of two protein structures docking of two protein structures that gives the best physical and geometric that gives the best physical and geometric complementaritycomplementarity

–– Simulation of metabolic pathwaysSimulation of metabolic pathways –– network simulation based on network simulation based on biomass conservation, kinetics equations, biomass conservation, kinetics equations, ……..

–– …………..


Molecular dynamics simulation

Metabolic pathway simulation


•• StatisticsStatistics--based computationbased computation –– discovering patterns in discovering patterns in unstructured dataunstructured data–– Computational methods for gene findingComputational methods for gene finding –– recognition of recognition of

distinguishing statistical features of coding regions from nondistinguishing statistical features of coding regions from non--coding coding regionsregions

–– Computational methods for promoter identificationComputational methods for promoter identification –– recognition of recognition of distinguishing features of promoter regions from non promoter distinguishing features of promoter regions from non promoter regionsregions

–– Computational methods for Computational methods for microarraymicroarray gene expression data gene expression data interpretationinterpretation –– identification of significant patterns in a data set, identification of significant patterns in a data set, which are possibly associated with a particular experimental which are possibly associated with a particular experimental conditioncondition

–– …………..


gene finding

gene chip data analysis

proteomic data analysis


•• BioinformaticBioinformatic support for modern biologysupport for modern biology–– Data storage, management, representation, and dissimilationData storage, management, representation, and dissimilation: : ----

modern biology generates great amount of datamodern biology generates great amount of data

–– Data interpretation for analytical instrumentsData interpretation for analytical instruments –– xx--ray ray crystallographic/NMR data interpretation, mass spec. data crystallographic/NMR data interpretation, mass spec. data interpretation, DNA sequencer data interpretation, interpretation, DNA sequencer data interpretation, ……..

–– Computational infrastructure for modern biologyComputational infrastructure for modern biology –– sample sample tracking systems, laboratory information management tracking systems, laboratory information management systems, systems, ……

–– …………....


databases


• data management/process; data mining; modeling; prediction; hypothesis formulation

engineering aspect

scientific aspect

bioinformatics

an indispensable part of biological science

genes, proteins, protein complexes, pathways, cells, organisms, eco-systems .............

What Can Biology Do for What Can Biology Do for Computational ScienceComputational Science

•• What computational science is to molecular What computational science is to molecular biology is like what mathematics has been to biology is like what mathematics has been to physicsphysics ............

Change of ParadigmChange of Paradigm•• The human genome sequencing project has led to fundamental The human genome sequencing project has led to fundamental

changes in how biological science is done!changes in how biological science is done!–– It represents biologyIt represents biology’’s first foray into s first foray into ‘‘big sciencebig science’’ –– ScienceScience, editorial, , editorial,

20032003

•• The coordinated efforts, funded by federal and private agencies,The coordinated efforts, funded by federal and private agencies, in in ““highhigh--throughputthroughput”” production of biological data beyond sequences production of biological data beyond sequences have fueled the rapid transition of biology from have fueled the rapid transition of biology from ““cottage industrycottage industry””science to science to ““big sciencebig science””–– functional genomic datafunctional genomic data–– structural genomic datastructural genomic data–– proteomic dataproteomic data–– haplotypehaplotype mappingmapping–– metabolonicmetabolonic datadata–– ……....

Change of ParadigmChange of Paradigm

high-throughput data production

data mining and interpretation

rational experimental

design

domain expertisemodels or hypotheses

specific data generation

Integrative BiologyIntegrative Biology

•• ““Howdy, want to do biology together?Howdy, want to do biology together?””

Examples of Integrated Computational Examples of Integrated Computational and Experimental Biologyand Experimental Biology

•• Structure prediction of Structure prediction of vitronectinvitronectin (a three(a three--domain protein) through domain protein) through combined computational and experimental workcombined computational and experimental work

•• XuXu lab did computational prediction of each of the domains of lab did computational prediction of each of the domains of vitronectinvitronectin

•• Peterson lab of UTK did NMR work to determine a few crossPeterson lab of UTK did NMR work to determine a few cross--links links between domainsbetween domains

•• Together we did the structure predictionTogether we did the structure prediction

We are carrying out similar work with Prestegard lab in a systematic manner


• Operon prediction in prokaryotic genomes

• Xu lab did operon prediction in cyanobacteria

•• PalenikPalenik lab of UCSD did experimental validation of some of the lab of UCSD did experimental validation of some of the predicted predicted opreonsopreons

•• Together we predicted Together we predicted opreonopreon prediction of prediction of cyanobacteriacyanobacteria at at genome scalegenome scale

Xu lab and Adams lab are carrying out similar studies on P. furiosus


•• Prediction of nitrogen assimilation pathway and its crossPrediction of nitrogen assimilation pathway and its cross--talk talk network with photosynthesis pathway in network with photosynthesis pathway in cyanobacteriacyanobacteria

•• XuXu lab did the initial prediction of component genes and possible lab did the initial prediction of component genes and possible interactions among the gene products in these pathway/networkinteractions among the gene products in these pathway/network

•• PalenikPalenik lab of UCSD did a number of lab of UCSD did a number of microarraymicroarray chips, knockout chips, knockout experiments and proteomic experiments, based on our initial experiments and proteomic experiments, based on our initial predictions predictions

•• Through a few iterations, we together have predicted the nitrogeThrough a few iterations, we together have predicted the nitrogen n assimilation pathway and its crossassimilation pathway and its cross--talk network with photosynthesistalk network with photosynthesis

CrossCross--talk Network between Nitrogen Assimilation and talk Network between Nitrogen Assimilation and PhotosynthesisPhotosynthesis

CO2

2-OG

NH4+

Glu

GS

Glu

Gln

Cyanate

Urease

NO3-

NO2-

NarB

NirA

Cyanase

Amt

Light

PsbA3, ApcF, PetF2, …PetH, Cytb6/f, CP43, …

RbcL, RbcS, Icd

Photosystem Calvin cycleATPNADPH+

DNA

Hypotheticalproteins

PII

NtcA regulonNon-ntcA regulon

…

Urea

GltS

Color codes:Shape codes:transportergeneprotein

Transcription factor

NtcAGO

GA

T

SomNutrients

Periplasmic membrane

Plasma membrane

SYNW0273

SYNW2289PetH

SYNY2460,2468,2469,2474

Rpod

Hypotheticalproteins

…

Other pathways

Cyanate

NO3

NrtP

UreaUrt

Cyn

NH4+ Glu

regulationtransformation/translocation


•• Understanding the mechanisms of alternative splicing in Understanding the mechanisms of alternative splicing in eukaryoteseukaryotes–– >80% of proteins in human cells are splicing variants>80% of proteins in human cells are splicing variants

•• Incorrect balance of different Incorrect balance of different isoformsisoforms in a cell could cause in a cell could cause diseases diseases –– cystic fibrosis (CF) cystic fibrosis (CF) –– spinal muscular atrophy (SMA)spinal muscular atrophy (SMA)–– FrontotemporalFrontotemporal dementiadementia


•• Our lab has been doing prediction of regulatory elements Our lab has been doing prediction of regulatory elements involved in regulation of alternative splicinginvolved in regulation of alternative splicing

““Systems BiologySystems Biology””

•• Integration of computational, experimental biology and highIntegration of computational, experimental biology and high--throughput probing technologies allows scientists to study throughput probing technologies allows scientists to study ““systemssystems””rather than individual elementsrather than individual elements

•• One argument is that a One argument is that a ““systemsystem”” is more than just the sum of all is more than just the sum of all individual parts; it could have emerging properties of its own, individual parts; it could have emerging properties of its own, based based on the complex (2on the complex (2--, 3, 3--, , ……., n., n--body) interactions among its partsbody) interactions among its parts

•• While systems emerging properties have been studied in other While systems emerging properties have been studied in other disciplines, such as stock market, not much has been discussed idisciplines, such as stock market, not much has been discussed in n the setting of biology the setting of biology


•• E. E. VoitVoit and colleagues of and colleagues of GatechGatech (Nature, 2005) recently carried out (Nature, 2005) recently carried out a a ““systems biologysystems biology”” study on the construction and modeling of study on the construction and modeling of sphingolipidsphingolipid metabolismmetabolism


•• Used previously published data to construct initial model of theUsed previously published data to construct initial model of thepathway modelpathway model

•• Refined the model using various experimental dataRefined the model using various experimental data

•• Validated the final pathway model through comparing dynamics Validated the final pathway model through comparing dynamics simulation results and timesimulation results and time--course data (metabolites) course data (metabolites)

•• Revealed a number of interesting features of the pathway, usefulRevealed a number of interesting features of the pathway, useful for for guiding further experimental designguiding further experimental design


•• …….. a .. a ““systemsystem”” is probably more than just the is probably more than just the sum of all individual parts; it could have sum of all individual parts; it could have emerging properties of its ownemerging properties of its own

•• Like physics, where Like physics, where general rules and lawsgeneral rules and laws are taught at the are taught at the start, start, biology will surely be presented to future generations of biology will surely be presented to future generations of students as a set of basic systems students as a set of basic systems ....... duplicated and ....... duplicated and adapted to a very wide range of cellular and adapted to a very wide range of cellular and organismicorganismicfunctions, functions, following basic evolutionary principles constrained following basic evolutionary principles constrained by Earthby Earth’’s geological history.s geological history.–– Temple SmithTemple Smith, , Current Topics in Computational Molecular BiologyCurrent Topics in Computational Molecular Biology

Look into the FutureLook into the Future

TakeTake--Home MessageHome Message•• MANY biological databases about functions and structures of MANY biological databases about functions and structures of

biomoleculesbiomolecules, processes and systems have been developed and , processes and systems have been developed and provided with powerful search enginesprovided with powerful search engines–– so much information is at the fingertips of biologistsso much information is at the fingertips of biologists–– why NOT to use them?why NOT to use them?

•• By fully combining experimental techniques with computation, By fully combining experimental techniques with computation, biologists can probably do much more than what they can do usingbiologists can probably do much more than what they can do usingexperiments (or computation) aloneexperiments (or computation) alone

•• More than ever, biological studies need infrastructure such as More than ever, biological studies need infrastructure such as bioinformaticsbioinformatics and and highhigh--throughput measurement technologythroughput measurement technology, to , to carry out complex systemcarry out complex system--level studieslevel studies

That is all, Folks.

THANK YOU!

the basics - computational systems biology labcsbl.bmb.uga.edu › ~xyn › compbio-intro.pdf · an...

Documents