基因芯片的数据分析

1[1][2,3]1231(differenceexpression,DE)2t1.1(foldchange,FC)[4]ratioratiocy3/cy5R/G0.5-2.0[5,6][7]210[8,9]1.2t(t-test)t[10]t2tt(regularizedt-test)()t-test[11]1.3(analysisofvariance,ANOVA)(ANOVA)F[12]mRNA[13]ABCD4ABCDBCD(post-hoccomparisons)[14]t-2[15,16]41(Bonferroni)CorrectedP-valueP-value×n(numberofgenesintest)P(0.05)2BonferroniStep-down(Holm)PCorrectedP-valueP-value×n(n-1/n-2……)P(0.05)3Westafall&Young2PPPPPPPP(0.05)4Benjamini&Hochberg4()()()()rrrrBAAABBΡΡΡ=Ρ3pPPCorrectedp-valuePvalue×(n/n-1)P0.053(family-wiseerrorrate,FWER)(falsediscoveryrate,FDR)(FDR)BonferroniWestfall&Young[17,18]1.4(nonparametricanalysis)“”t-[19]t-(nonparametrict-test)Wilcoxon(Wilcoxonranksuntest)[20](empiricalBayesmethod)[21](significanceanalysisofmicroarray,SAM)[22](themixturemodelmethod,MMM)[23]()1.5(RegressionAnalysis)()[24,25]Lietal[26](Cox)Huangetal[27]2(clusteringanalysis)()4[28]2.1(hierarchicalclustering,HCL)nn[29][30][31,32]2.2K-(k-meansclustering,KMC)K-K-[33]D'ambrosioetal[34]K-103K-[35]2.3(self-organizingmapclustering,SOM)2[36]Covelletal[37]1480%2.4(two-wayclustering,TWC)(one-wayclustering)(two-wayclustering)[38](geneshaving,GS)(plaidmodels)1221()diiABnixy==⎡⎤∑−⎢⎥⎣⎦5Hastieetal[39]BJiangetal[40]213105[41,42]2.5(unsupervised)(supervised)Wangetal[43]K-Herreroetal[44]3(discriminantanalysis)()[45,46]3.1(fisherdiscriminantanalysisFDA)[47]Choetal[48]Dangondetal[49]3.2(bayesiannetworks)(causalnetworks)[50]Imotoetal[51]6Kimetal[52]3.3(supportvectormachines,SVMs)()()“”[53-55]Williamsetal[56]271323.4(decisiontrees)[57]Dettlingetal[58]Middendorfetal[59]3.5(artificialneuralnetwork,ANN)ANN(“”“”)“”()[60]O'Neilletal[61]Sawaetal[62]344.1(principalcomponentanalysis,PCA)[63,64]——7[65]Crescenzietal[66]6013755Yeungetal[67]PCA1ReimersM.Statisticalanalysisofmicroarraydata.AddictBiol.2005;10:23-35[PMID:15849016]2HacklH,CaboFS,SturnA,WolkenhauerO,TrajanoskiZ.AnalysisofDNAmicroarraydata.CurrTopMedChem.2004;4:1357-1370[PMID:15379650]3LeungYF,CavalieriD.FundamentalsofcDNAmicroarraydataanalysis.TrendsGenet.2003;19:649-659[PMID:14585617]4GerholdD,LuM,XuJ,AustinC,CaskeyCT,RushmoreT.MonitoringexpressionofgenesinvolvedindrugmetabolismandtoxicologyusingDNAmicroarrays.PhysiolGenomics,2001;5:161-70[PMID:11328961]5MutchDM,BergerA,MansourianR,RytzA,RobertsMA.Thelimitfoldchangemodel:apracticalapproachforselectingdifferentiallyexpressedgenesfrommicroarraydata.BMCBioinformatics,2002;3:17[PMID:12095422]6YangIV,ChenE,HassemanJP,LiangW,FrankBC,WangS,SharovV,SaeedAI,WhiteJ,LiJ,LeeNH,YeatmanTJ,QuackenbushJ.Withinthefold:assessingdifferentialexpressionmeasuresandreproducibilityinmicroarrayassays.GenomeBiol,2002;3(11):research0062[PMID:12429061]7BlackMA,DoergeRW.Calculationoftheminimumnumberofreplicatespotsrequiredfordetectionofsignificantgeneexpressionfoldchangeinmicroarrayexperiments.Bioinformatics,2002;18:1609-1616[PMID:12490445]8CuiX,ChurchillGA.StatisticaltestsfordifferentialexpressionincDNAmicroarrayexperiments.8GenomeBiol,2003;4:210[PMID:12702200]9DraghiciS.Statisticalintelligence:effectiveanalysisofhigh-densitymicroarraydata.DrugDiscovToday,2002;7:S55-63[PMID:12047881]10BaldiP,LongAD.ABayesianframeworkfortheanalysisofmicroarrayexpressiondata:regularizedt-testandstatisticalinferencesofgenechanges.Bioinformatics,2001;17:509-519[PMID:11395427]11LongAD,MangalamHJ,ChanBY,TolleriL,HatfieldGW,BaldiP.ImprovedstatisticalinferencefromDNAmicroarraydatausinganalysisofvarianceandaBayesianstatisticalframework.AnalysisofglobalgeneexpressioninEscherichiacoliK12.JBiolChem,2001;276:19937-19944[PMID:11259426]12PavlidisP.UsingANOVAforgeneselectionfrommicroarraystudiesofthenervoussystem.Methods,2003;31:282-289[PMID:14597312]13HatfieldGW,HungSP,BaldiP.DifferentialanalysisofDNAmicroarraygeneexpressiondata.MolMicrobiol,2003;47:871-877[PMID:12581345]14PanKH,LihCJ,CohenSN.AnalysisofDNAmicroarraysusingalgorithmsthatemployrule-basedexpertknowledge.ProcNatlAcadSciUSA,2002;99:2118-2123[PMID:11854507]15AubertJ,Bar-HenA,DaudinJ,RobinS.Correction:DeterminationofthedifferentiallyexpressedgenesinmicroarrayexperimentsusinglocalFDR.BMCBioinformatics,2005;6:42[PMID:15745450]16PawitanY,MurthyKR,MichielsS,PlonerA.Biasintheestimationoffalsediscoveryrateinmicroarraystudies.Bioinformatics.2005;21:3865-3872[PMID:16105901]17PawitanY,MichielsS,KoscielnyS,GusnantoA,PlonerA.Falsediscoveryrate,sensitivityandsamplesizeformicroarraystudies.Bioinformatics.2005;21:3017-3024[PMID:15840707]18GrantGR,LiuJ,StoeckertCJJr.Apracticalfalsediscoveryrateapproachtoidentifyingpatternsofdifferentialexpressioninmicroarraydata.Bioinformatics.2005;21:2684-2690[PMID:15797908]19ZhaoY,PanW.Modifiednonparametricapproachestodetectingdifferentiallyexpressedgenesinreplicatedmicroarrayexperiments.Bioinformatics,2003;19:1046-1054[PMID:12801864]20TroyanskayaOG,GarberME,BrownPO,BotsteinD,AltmanRB.Nonparametricmethodsforidentifyingdifferentiallyexpressedgenesinmicroarraydata.Bioinformatics,92002;18:1454-1461[PMID:12424116]21EfronB,TibshiraniR.Empiricalbayesmethodsandfalsediscoveryratesformicroarrays.GenetEpidemiol,2002;23:70-86[PMID:12112249]2