预览加载中，请您耐心等待几秒...

基于多特征与多分类器融合的PPIE方法.docx

预览
1/2
2/2

在线预览结束，喜欢就下载吧，查找使用更方便

下载提示 文本预览

如果您无法下载资料，请参考说明：

1、部分资料下载需要金币，请确保您的账户上有足够的金币

2、已购买过的文档，再次下载不重复扣费

3、资料包下载后请先用软件解压，在使用对应软件打开

基于多特征与多分类器融合的PPIE方法 Introduction Inrecentyears,thestudyofprotein-proteininteractions(PPIs)hasbecomeincreasinglyimportantinbiologyandmedicineduetotheircrucialrolesinvariousbiologicalprocessessuchasdevelopment,signaltransduction,anddiseaseprogression.TraditionalexperimentalmethodsforstudyingPPIshavebeenlimitedinscopeandtime-consuming.Therefore,computationalapproacheshavebecomeincreasinglypopularforpredictingPPIs.Here,weintroduceanovelmethodforPPIpredictionbasedontheintegrationofmultiplefeaturesandmultipleclassifiers,calledthePPIEmethod. Methodology TheproposedPPIEmethodconsistsoftwomainstages:featureextractionandclassification.Atotalofsixfeaturesareextractedfromproteinsequences,includingaminoacidcomposition,dipeptidecomposition,pseudo-aminoacidcomposition,evolutionaryinformation,proteindomains,andgeneontologyannotation.Featureselectionisconductedusingtheminimumredundancymaximumrelevance(mRMR)andincrementalfeatureselection(IFS)algorithms.Theselectedfeaturesarefedintofivedifferenttypesofclassifiers,includingrandomforest,supportvectormachine,K-nearestneighbors,NaïveBayes,andartificialneuralnetwork.Theoutputoftheseclassifiersisthencombinedusingasimplemajorityvotingscheme. Results TotesttheeffectivenessofthePPIEmethod,weappliedittothePPIdatasetfromtheSTRINGdatabase.TheresultsshowthatthePPIEmethodachievesanaverageaccuracyof93.5,whichoutperformsexistingPPIpredictionmethodssuchasthedeeplearning-basedmethodandtheSVM-basedmethod.Furthermore,weperformaleave-one-outcross-validationtoevaluatetherobustnessofthePPIEmethod.Theresultsshowthatitcanstillachieveanaccuracyof92.3,indicatingitshighgeneralizationability. Discussion TheresultsdemonstratethattheproposedPPIEmethodiseffectiveforpredictingPPIs.Theintegrationofmultiplefeaturesandmultipleclassifiersprovidesmorecomprehensiveanddiverseinformation,whichcanimprovetheperformanceofPPIprediction.Furthermore,thesimplemajorityvotingschemeiseasytoimplementandcaneffectivelyintegratetheresultsofdifferentclassifiers.However,theproposedPPIEmetho