资源描述：
《利用PLS_VIP方法筛选差异表达基因_英文_韩放.pdf》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。
1、北京大学学报(自然科学版),第45卷,第1期,2009年1月ActaScientiarumNaturaliumUniversitatisPekinensis,Vol.45,No.1(Jan.2009)SearchingforDifferentiallyExpressedGenesbyPLS-VIPMethod1111,2,HANFang,WUJingchen,XUJiangfeng,DENGMinghua11SchoolofMathematicalScience,PekingUniversity,Be
2、ijing100871;21CenterforTheoreticalBiology,PekingUniversity,Beijing100871;CorrespondingAuthor,E-mail:dengmh@math.pku.edu.cnAbstractAnewapproachcalledPLS-VIPbasedonvariablespimportanceofprojection(VIP)fordetectingdifferentiallyexpressedgenesisproposed.Iti
3、sanewmethodtosumupthecontributionofcomponentsbasedonPartialLeastSquaremethod(PLS).Asittakesthecorrelationofdifferentgenesintoconsideration,itismoresuitablethanthoseclassificationmethodsbasedonreviewingdistinctgenesseparately.Theeffectofgenesextractedbyo
4、rdinaryPLS,discriminantPLSandproposedmethodtoclassifymult-iclasstumorsiscompared.Resultsshowthattheerrorrateofthenewmethodisobviouslylowerthantheothertwomethodsinmostcases.Keywordsmicroarray;differentiallyexpressedgenes;reductionofdimension;PLS-VIP利用PLS
5、-VIP方法筛选差异表达基因1111,2,韩放吴晶辰徐江峰邓明华11北京大学数学科学学院,北京100871;21北京大学理论生物学中心,北京100871;通讯作者,E-mail:dengmh@math.pku.edu.cn摘要提出一种基于变量权重寻找差异表达基因的新方法。该方法的最终目的是从微阵列数据中抽取出核心变量(基因)。将该种方法抽取出的差异表达基因判别样本的能力和普通的PLS方法以及判别最小二乘方法进行比较,结果表明该方法的错误率明显低于其他两种传统方法。因此,PLS-VIP方法是一种较为合适
6、的抽取差异表达基因并判别样本的方法。关键词微阵列;差异表达基因;降维;PLS-VIP中图分类号O213Microarrayshavebecomeincreasinglycommonin(DE)areoftenreferredtoasclinicalmarkers.biologicalandmedicalresearch.TheyenabletheIdentificationofclinicalmarkersmayleadtoimprovedsimultaneousstudyofthousandsofg
7、enesandafforddiagnosisandtreatmentguidance,earlydiseasedetectionunprecedentedabilitytoprovidegeneexpressionandclinicaloutcomesprediction.informationonawholegenomelevel.WiththeadventofThemostcommonlyusedtoolsforidentificationofDNAmicroarraytechnology,itb
8、ecamepossibletodetectdifferentiallyexpressedgenesinclude-ttest,logisticdifferentiallyexpressedgenesonagenome-widescale.Aregressionandlinearregressionanalysis,someofwhichmajorquestioninmicroarraystudiesishowtoselectgenesoftenrequi
