熊江辉_用网络药理学研究药物组合_中药研究潜在应用

InsilicoGeneticNetworkModelsforPre-clinicalDrugPrioritizationJianghuiXiong熊江辉[e-mail]://:Genomics,ProteomicsandMetabolomicsinDrugDiscoveryandTargetValidation点击获取该研究发表的论文全文（PLoSONE）：•Networkcouldbedrugtarget•JianghuiXiongetc.,Pre-clinicaldrugprioritizationviaprognosis-guidedgeneticinteractionnetworks.PloSONE2010•“Formorethanadecade,scientistsinsystemsbiologyhavepromisedthatrealbreakthroughingeneticmedicinewillcomewhenwestopmappingindividualgenestophenotypesandinsteadstartlookingatinteractingnetworks.Yet,notmuchhashappened.Thefieldisstillstrugglingtodefinerelevantnetworksandtointerpretdataintermsofthosenetworks.ThepaperbyXiongetaladdsconsiderablytotheprogressofnetwork-basedgeneticmedicine.Itishighlyrelevant,originalandinteresting.”OncologyDrugDevelopmentOneofmostchallengingscientificproblemsWhat’swrongwithourcancermodels?NATUREREVIEWSDRUGDISCOVERY,2005Thecurrentmodelsusedforpre-clinicaldrugtestingDoNOTaccuratelypredicthownewtreatmentswillactinclinicaltrials–Heterogeneityinpatientpopulations–UnpredictablephysiologyWhat’swrongwithourcancermodels?NATUREREVIEWSDRUGDISCOVERY,2005What’swrongwithourDiseaseModelsDrugDiscovTodayDisModels,2008?Hypothesis–Consideringgenenetworksassociatedwithcanceroutcomeinheterogeneouspatientpopulations–Thedifficultyofidentifyeffectivecancercures(asevidencedbydrugresistance)maybeaconsequenceoftherobustnessofthisnetwork–Network(robustness)asdrugtargetPre-clinicalinsilicoCancerModelsforDrugactionstudy–Incorporatingheterogeneityandinvivophysiologyinformation,whichMISSINGinpre-clinicalcancermodelsOurproposal12SOD(SynergisticOutcomeDetermination)BadOutcomeGenePairGeneAGeneBBadOutcomeGenePairGeneAGeneBGoodOutcomeGenePairGeneAGeneBGoodOutcomeGenePairGeneAGeneBSynergisticallyInferedNexus(SIN)𝑆𝑦𝑛𝐺1,𝐺2;𝐶=𝐼𝐺1,𝐺2;𝐶−𝐼𝐺1;𝐶+𝐼(𝐺2;𝐶)𝐼𝑋;𝑌=𝑝𝑥,𝑦log2𝑝(𝑥,𝑦)𝑝𝑥𝑝(𝑦)𝑦𝑥SOD(SynergisticOutcomeDetermination)vsSyntheticLethalityFeaturecomparedSODSyntheticLethalityPhenotypeSurvivaloutcomeofindividualpatientCelldeath/growthLevelIndividualCellDataAccessibleHumanpopulation(viacomputation)Yeast(SGA);Humancelllines;HumanpopulationThepipelineProteinNetworkPathwayGeneOntologymiRNATargetGeneGeneModuleDatabasePrognosticDataInter-ModuleCooperationNetwork1CompoundinvitroscreendataNCI60panelgeneExpressiondataGeneSignatureforCompoundSensitivityCompoundrouteofactionAnalysisPerturbationIndex(PI)Benchmarking/ValidationDrugCombinationSimulation23WhatisGeneModule?AndWhyWeuseitinsteadofthesinglegenes?GeneModule:agroupofgeneswhichsharesimilarfunctionabcdefGeneModule1xAsinglegene……GeneModule2GeneModule3GeneModule4GeneModulen……GeneModule:robust/reproduciblefeaturesratherthansinglegenesGeneModuleDatabaseProteinNetworkPathwayGeneOntologymiRNATargetGeneGeneModuleDatabaseChangHYetal.PNAS2005;102:3738-3743a“woundresponse”geneexpressionsignatureinpredictingbreastcancerprogression•Naturalpopulation-Heterogeneity•Tumortissue-Microenvironmentreflection•Finalpointphenotype-Survivaltime•Comprehensivegenomiccharacterization•LargeDataSetPrognosisData--dataassociatedgeneexpressionwithpatients’prognosisBenefitofPrognosisDataPrognosisDataInstanceGeneModuleDatabaseModule-modulecooperationnetworkK12...ForeachgeneQueryModuleAGeneKSINGenelistGene2SINGenelistWholeGeneSet(Mgenes)Gene-GeneSINanalysisCandidateModuleBCandidateModuleCGene1SINGenelistSINAQueryModuleASINACandidateModuleCSINACooperationModuleCSINACandidateModuleBSINAQueryModuleACooperationModuleBPoolTogetherInter-ModuleCooperationNetwork(IMCN)forlungcancersuggeststhatthenetworkrobustnesshighlydependentongatekeepermodulesGatekeeperModuleCheckpointModuleCharacterizationoftheInter-ModuleCooperationNetwork(IMCN)‘Gatekeeper’modulesforlungcancer(NSCLC)3MajorBiologicalThemeContributionofvariousevidencesourcesforgenemoduledefinitionComparinggenetic(somaticmutation)andepigenetic(DNAmethylation)aberrationrate(intumorvs.normal)oftwotypesofmodulesTop10%or20%ofgeneswhichhighlyused(i.e.onegeneinvolvedinmultiplegenemodules)asrepresentativeofeachtypesofmodulesCompoundactiononcellsNCI60invitroDrugscreenProjectConnectivityMAPUsePerturbationIndex(PI)toquantifyDrugactionHypothesis–Todisrupt/perturbcancernetwork,thekeytosuccessistosimultaneouslyperturbsthecorrespondinggatekeepermoduleswiththecheckpointmodules•Hi--thenumberofhitsbycompoundc•Li--theactivelinks(i.e.linksinwhichbothsourcenodeandtargetnodearematchedbycompoundc)•N--thenumberofgatekeepermodulesBenchmarkingforpre-clinicaldrugprioritizing•Whytest?–Assessthepotentialapplicationforprioritizingcompoundsforclinicaltrials,basedontheinformationavailableinpre-clinicalstage•‘StandardAgentDatabase’–OriginallycreatedbyBoyd[29]andultimatelyfinalizedbytheNCI–CompoundswhichhavebeensubmittedtotheFDAforreviewasaNewDrugApplication–ORcompoundsthathavereachedaparticularhighstageofinterestattheNCI•Successfuldruglist-