Applied Statistics for Network Biology

479 Pages·2011·7.802 MB·English

by  Dehmer M., et al. (eds.)

Edited by
Matthias Dehmer,
Frank Emmert-Streib
Armin Graber, and
Armindo Salvador
Applied Statistics
for Network Biology
Related Titles
Emmert-Streib, F.,Dehmer, M.(eds.)
Medical Biostatistics for Complex Diseases
2010
ISBN:978-3-527-32585-6
Dehmer,M.,Emmert-Streib, F.(eds.)
Analysis of Complex Networks
FromBiologytoLinguistics
2009
ISBN:978-3-527-32345-6
Emmert-Streib, F.,Dehmer, M.(eds.)
Analysis of Microarray Data
ANetwork-BasedApproach
2008
ISBN:978-3-527-31822-3
Junker,B. H.,Schreiber,F.
Analysis of Biological Networks
2008
ISBN:978-0-470-04144-4
Stolovitzky,G., Califano,A.(eds.)
Reverse Engineering Biological Networks
OpportunitiesandChallengesinComputationalMethodsforPathwayInference
2007
ISBN:978-1-57331-689-7
Quantitative and Network Biology
Series Editors M. Dehmer and F. Emmert-Streib
Volume 1
Applied Statistics for Network Biology
Methods in Systems Biology
Edited by
Matthias Dehmer, Frank Emmert-Streib, Armin Graber,
and Armindo Salvador
TheEditors LimitofLiability/DisclaimerofWarranty:Whilethe
publisherandauthorhaveusedtheirbesteffortsin
MatthiasDehmer preparingthisbook,theymakenorepresentations
UMIT orwarrantieswithrespecttotheaccuracyor
InstituteforBioinformatics completenessofthecontentsofthisbookand
andTranslationalResearch specificallydisclaimanyimpliedwarrantiesof
EduardWallnöferZentrum1 merchantabilityorfitnessforaparticularpurpose.
6060Hall,Tyrol Nowarrantycanbecreatedorextendedbysales
Austria representativesorwrittensalesmaterials.TheAdvice
andstrategiescontainedhereinmaynotbesuitable
foryoursituation.Youshouldconsultwitha
FrankEmmert-Streib
professionalwhereappropriate.Neitherthe
Queen’sUniversityBelfast
publishernorauthorsshallbeliableforanylossof
CenterforCancerResearchandCellBiology
profitoranyothercommercialdamages,including
97,LisburnRoad
butnotlimitedtospecial,incidental,consequential,
BelfastBT97BL
orotherdamages.
UnitedKingdom
LibraryofCongressCardNo.: appliedfor
ArminGraber
UMIT BritishLibraryCataloguing-in-PublicationData
InstituteforBioinformatics Acataloguerecordforthisbookisavailablefromthe
andTranslationalResearch BritishLibrary.
EduardWallnöferZentrum1
6060Hall,Tyrol Bibliographicinformationpublishedby
Austria theDeutscheNationalbibliothek
TheDeutscheNationalbibliothekliststhis
and
publicationintheDeutscheNationalbibliografie;
NovartisPharmaceuticalsCorporation detailedbibliographicdataareavailableonthe
OncologyBiomarkersandImaging Internetathttp://dnb.d-nb.de.
OneHealthPlaza
EastHanover,NJ07936 #2011Wiley-VCHVerlag&Co.KGaA,
USA Boschstr.12,69469Weinheim,Germany
ArmindoSalvador Wiley-BlackwellisanimprintofJohnWiley&Sons,
formedbythemergerofWiley’sglobalScientific,
UniversityofCoimbra
Technical,andMedicalbusinesswithBlackwell
CenterforNeuroscienceand
Publishing.
CellBiology,DepartmentofChemistry
3004-535Coimbra
Allrightsreserved(includingthoseoftranslationinto
Portugal
otherlanguages).Nopartofthisbookmaybe
reproducedinanyform–byphotoprinting,
Composition ThomsonDigital,Noida,India microfilm,oranyothermeans–nortransmittedor
PrintingandBinding betz-druckGmbH,Darmstadt translatedintoamachinelanguagewithoutwritten
CoverDesign AdamDesign,Weinheim permissionfromthepublishers.Registerednames,
trademarks,etc.usedinthisbook,evenwhennot
specificallymarkedassuch,arenottobeconsidered
unprotectedbylaw.
PrintedintheFederalRepublicofGermany
Printedonacid-freepaper
ISBN:978-3-527-32750-8
V
Contents
Preface XVII
List of Contributors XIX
PartOne Modeling,Simulation,andMeaningofGeneNetworks 1
1 NetworkAnalysistoInterpretComplexPhenotypes 3
HongYu,JialiangHuang,WeiZhang,andJing-DongJ.Han
1.1 Introduction 3
1.2 IdentificationofImportantGenesbasedonNetwork
Topologies 5
1.2.1 Degree 5
1.2.2 Betweenness 6
1.2.3 NetworkMotifs 6
1.2.4 HierarchicalStructure 7
1.3 InferringInformationfromKnownNetworks 8
1.3.1 UnderstandingBiologicalFunctionsbasedon
NetworkModularity 8
1.3.2 InferringFunctionalRelationshipsandNovelFunctional
GenesThroughNetworks 8
1.3.3 UnravelingTranscriptionalRegulationsfromExpression
DatathroughTranscriptionalNetworks 9
1.3.4 ExtractingthePathway-LinkedRegulatorsandEffectors
basedonNetworkFlows 10
1.4 Conclusions 10
References 11
2 StochasticModelingofGeneRegulatoryNetworks 13
TianhaiTian
2.1 Introduction 13
2.2 DiscreteStochasticSimulationMethods 14
2.2.1 SSA 15
2.2.2 Acceleratingt-LeapMethods 16
2.2.3 LangevinApproach 19
VI Contents
2.3 DiscreteStochasticModeling 20
2.3.1 StochasticModelingMethod 20
2.3.2 ToggleSwitchwiththeSOSPathway 22
2.3.3 OtherModelsfortheGeneticToggleSwitch 24
2.4 ContinuousStochasticModeling 26
2.4.1 DeterministicModelsforthelPhageNetwork 26
2.4.2 StochasticModelsforExternalNoise 28
2.4.3 DeterministicModelswithThresholdValues 29
2.4.4 StochasticSwitching 30
2.5 StochasticModelsforBothInternalandExternal
Noise 31
2.5.1 StochasticModelsforMicroarrayGene
ExpressionData 33
2.6 Conclusions 34
References 35
3 ModelingExpressionQuantitativeTraitLociinMultiple
Populations 39
Ching-LinHsiaoandCathyS. J.Fann
3.1 Introduction 39
3.1.1 DataStructureineQTLStudies 39
3.1.2 CurrenteQTLStudies 40
3.1.2.1 eQTLStudiesinaSingleHumanPopulation 40
3.1.2.2 eQTLStudiesinMultipleHumanPopulations 43
3.1.3 AnIllustratedExample 45
3.1.4 Notations 46
3.2 IGMMethod 47
3.2.1 ModelingSNP–GEAssociationinaSinglePopulation 47
3.2.2 IntegratingHypothesestoIdentifyCommoneQTL 48
3.2.3 ApplyingtheIGMMethodtoHapMapData 48
3.2.3.1 CharacterizingPutativeeQTLIdentifiedbytheIGM 49
3.3 CTWM 50
3.3.1 ModelingSNP–GEAssociationinPooledData
byCTWM 50
3.3.2 ApplyingCTWMtoHapMapData 52
3.3.2.1 CharacterizingPutativeeQTLIdentifiedbyCTWM 52
3.3.2.2 JustificationofModelAssumptions 53
3.4 CTWM-GSMethod 53
3.4.1 SolvingNormalEquationsinCTWM 54
3.4.2 EstimatorsofBDandGS 55
3.4.3 TestingBDandGS 56
3.4.4 ApplyingCTWM-GStoHapMapData 56
3.4.4.1 ApplyingtheGStoPopulationStudies 57
3.5 Discussion 60
References 61
Contents VII
PartTwo InferenceofGeneNetworks 67
4 TranscriptionalNetworkInferenceBasedon
InformationTheory 69
PatrickE.Meyer,CatharinaOlsen,andGianlucaBontempi
4.1 Introduction 69
4.1.1 Notation 69
4.1.2 Formalization 71
4.1.3 PerformanceMeasuresinUndirectedNetwork
Inference 72
4.1.3.1 Precision–Recall(PR)Curves 73
4.1.3.2 F-Scores 74
4.1.4 CausalSubsetSelection 74
4.2 InferenceBasedonConditionalMutual
Information 76
4.2.1 Constraint-BasedMethods 77
4.2.2 ApproximatedConditionalMutualInformation 78
4.2.3 VariableSelectionAlgorithms 78
4.3 InferenceBasedonPairwiseMutualInformation 80
4.3.1 RelevanceNetwork(RELNET) 80
4.3.2 ContextLikelihoodofRelatedness(CLR) 81
4.3.3 Chow–LiuTree 81
4.3.4 AlgorithmfortheReconstructionofAccurateCellular
Networks(ARACNE) 82
4.3.5 MinimumRedundancyNetworks(MRNET) 83
4.4 ArcOrientation 84
4.4.1 AssessingArcOrientationMethods 86
4.5 Conclusions 87
References 87
5 ElucidationofGeneralandCondition-DependentGenePathways
UsingMixtureModelsandBayesianNetworks 91
SandraRodriguez-ZasandYounheeKo
5.1 Introduction 91
5.2 Methodology 92
5.2.1 NetworkLearningAlgorithms:Frequentist-andBayesian
MCMC-basedalgorithms 93
5.3 Applications 95
5.3.1 ElucidationofGeneNetworks 95
5.3.2 DiscoveryofCondition-DependentGene
Relationships 96
5.3.3 MCMCMixtureBayesianNetwork 99
5.3.4 ComputationalConsiderations 101
5.4 Conclusions 101
References 102
VIII Contents
6 MultiscaleNetworkReconstructionfromGeneExpression
Measurements:Correlations,Perturbations,and‘‘APriori
BiologicalKnowledge’’ 105
DanielRemondiniandGastoneCastellani
6.1 Introduction 105
6.1.1 ComplexNetworks 105
6.1.2 GeneInteractionNetworksfromGeneExpression
Measurements 107
6.2 ‘‘PerturbationMethod’’ 108
6.3 NetworkReconstructionbytheCorrelationMethod
fromTime-SeriesGeneExpressionData 109
6.4 NetworkReconstructionfromGeneExpressionDataby
APrioriBiologicalKnowledge 110
6.5 ExamplesandMethodsofCorrelationNetworkAnalysis
onTime-SeriesData 112
6.6 ExamplesandMethodsforPathwayNetworkAnalysis 117
6.6.1 GeneSelectionandPathwayGrouping 118
6.6.2 PathwaySignificanceandPathwayNetwork 118
6.6.3 Results 121
6.7 Discussion 126
6.8 Conclusions 127
References 128
7 GeneRegulatoryNetworksInference:Combininga
GeneticProgrammingandH1FilteringApproach 133
LijunQian,HaixinWang,andXiangfangLi
7.1 Introduction 133
7.2 Background 134
7.2.1 NoiseinGeneExpression 134
7.2.2 ModelingofGeneRegulatoryNetworkswith
Noise 136
7.2.2.1 BooleanNetworksModelwithNoise 136
7.2.2.2 BayesianNetworksModelwithNoise 136
7.2.2.3 LinearAdditiveRegulationModelwithNoise 137
7.2.2.4 NeuralNetworksModelwithNoise 137
7.2.3 ProposedNonlinearODEModelwithNoise 138
7.3 MethodologyforIdentificationandAlgorithm
Description 139
7.3.1 GP 140
7.3.2 H1Filter 142
7.4 SimulationEvaluation 144
7.4.1 SyntheticData 144
7.4.2 MicroarrayData 145
7.5 Conclusions 146
References 150
Contents IX
8 ComputationalReconstructionofProteinInteractionNetworks 155
KonradMönks,IrmgardMühlberger,AndreasBernthaler,RaulFechete,
PaulPerco,RudolfFreund,ArnoLukas,andBerndMayer
8.1 Introduction 155
8.1.1 SelectingRelevantFeaturesfromOmicsProfiles 156
8.1.2 AnalyzingOmicsDataonaNetworkLevel 157
8.2 ProteinInteractionNetworks 159
8.2.1 NetworkCategories 159
8.2.1.1 MetabolicNetworks 159
8.2.1.2 ParalogNetworks 160
8.2.1.3 PhysicalInteractionNetworks 160
8.2.2 ParametersforProteinAnnotation 161
8.2.2.1 GeneExpressionProfiles 161
8.2.2.2 SubcellularLocation 161
8.2.2.3 GeneAnnotation 161
8.2.2.4 TranscriptionFactors 162
8.2.2.5 MicroRNA 162
8.2.2.6 Pathways 162
8.2.3 DataPreparation 163
8.2.3.1 IntegrationofDataSources 163
8.2.3.2 ObtainingEdgeWeights 164
8.2.3.3 DataCompleteness 166
8.2.3.4 DataNormalization 166
8.2.4 DerivingModels 166
8.2.4.1 BasicConsiderations 167
8.2.4.2 ChoosinganAlgorithm 169
8.2.5 ValidationProcedures 169
8.2.5.1 ModelPerformanceEvaluation 169
8.2.5.2 NetworkStructureAssessment 170
8.3 CharacterizationofComputedNetworks 171
8.3.1 EvaluationoftheSpecificProtein–ProteinInteractions 171
8.3.2 ApplicationoftheSpecificProtein–ProteinInteractions 175
8.4 Conclusions 177
References 178
PartThree AnalysisofGeneNetworks 181
9 WhatiftheFitisUnfit?CriteriaforBiologicalSystemsEstimation
BeyondResidualErrors 183
EberhardO.Voit
9.1 Introduction 183
9.2 ModelDesign 184
9.3 ConceptsandChallengesofParameterEstimation 187
9.3.1 TypicalParameterEstimationProblems 190
9.3.1.1 DataFitisUnacceptable 190