预览加载中，请您耐心等待几秒...

基于网络的癌症驱动基因预测方法的性能评估.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

基于网络的癌症驱动基因预测方法的性能评估 Title:PerformanceEvaluationofNetwork-basedCancerDriverGenePredictionMethods Introduction: Cancerisacomplexandheterogeneousdiseasethatarisesduetotheaccumulationofgeneticmutations.Identifyingdrivergenes,whichplayacrucialroleincancerinitiationandprogression,isessentialforunderstandingtheunderlyingmechanismsanddevelopingeffectivetargetedtherapies.Withtheadventofhigh-throughputsequencingtechnologies,large-scalegenomicdatasetshavebecomeavailable,enablingthedevelopmentofcomputationalmethodsforpredictingcancerdrivergenes.Inrecentyears,network-basedmethodshavegainedpopularityduetotheirabilitytocapturetheintrinsicrelationshipsbetweengenesandtheirinvolvementincancerpathways.Thispaperaimstoevaluatetheperformanceofnetwork-basedcancerdrivergenepredictionmethods. Methods: Toevaluatetheperformanceofnetwork-basedcancerdrivergenepredictionmethods,weselectedasetofwell-establishedmethodsthatutilizenetworkinformation,suchasprotein-proteininteractionnetworksorgeneco-expressionnetworks.ThesemethodsincludebutarenotlimitedtoHotNet2,DawnRank,OncodriveFM,andMUFFINN.Weusedabenchmarkdatasetconsistingofknowncancerdrivergenesandrandomlyselectednon-drivergenes.Performancemetricssuchasprecision,recall,F1score,andareaunderthereceiveroperatingcharacteristiccurve(AUROC)werecalculatedforeachmethod. Results: Theresultsofourperformanceevaluationrevealedvaryingdegreesofaccuracyamongthenetwork-basedcancerdrivergenepredictionmethods.HotNet2achievedthehighestprecision,recall,andF1score,indicatingitsabilitytoaccuratelypredictdrivergenes.DawnRankandOncodriveFMalsodemonstratedrelativelyhighperformance,whileMUFFINNexhibitedloweraccuracy.AUROCanalysisfurtherconfirmedthesuperiorperformanceofHotNet2indiscriminatingbetweendrivergenesandnon-drivergenes. Discussion: ThehighperformanceofHotNet2canbeattributedtoitsabilitytoidentifysignificantlyalteredsubnetworksincancersamples,whichhelpsincapturingrecurrentlyperturbedpathways.However,itisimportanttonotethattheperformanceofthesemethodsmayvarydependin