基于DEA理論的新陳代謝網(wǎng)絡(luò)中調(diào)控結(jié)構(gòu)的代謝物濃度預(yù)測.pdf

1、基于DEA理論的新陳代謝網(wǎng)絡(luò)中調(diào)控結(jié)構(gòu)的代謝物濃度預(yù)測MetaboliteConcentrationPredictioninRegulatoryArchitecturesofMetabolicReactionNetworksusingDEA碩『j學(xué)位論史學(xué)科專業(yè)：重簦堂皇控劍途研究生：武童查指導(dǎo)教師：墊馥蓮熬援天津大學(xué)理學(xué)院二零零八年五月ABSTRACTAsanonparametricmethod，DataEnvelopmentAnal

2、ysis(DEA)isappliedtomanyfieldsBasedonmetabolicnetworkanalysisinbiochemicalsystem，thisdissertationstudiesthemodelsandmethodstopredictthemetaboliteconcentrationThisdissertationconsistsoftwopartsFirstamodifiedC2GS2modelisus

3、edtopredictmetaboliteconcentrationwithdifferentregulatorystructuresThenumberofinequalityconstraintsinthemodifiedmodelisdecreased，butitsefficiencyismaintainedThreegroupsofdatawerecheckedusingDEAmodelsInthe1Stand2ndgroups，

4、theenzymeschangebetween士20％and500％andbetween410％oftheirreferencevalues，withoutmodificationsintheregulatorystructureInthe3rdgroup，theenzymeschangebetween士20％and500％oftheirreferencevalues，andthenumberOfbiochemicalreactions

5、isdecreasedThepredictedvaluesarevalidforthe1stand2ndgroups，andtheircombination，respectivelyButthemodelfailsforthedataconsistingofallthethreegroupsItisconcludedthatthemodelisonlyvalidfordatafromreactionnetworkswiththesame

6、regulatorystructure，andthischaracteristiccanbeused雒areferencetodistinguishmetaboliteswithdifferentregulatorystructuresSecond，byclassifyingthedecisionmakingunits(DMUs)，thepredictionefficiencyofmetaboliteswiththesameregula

7、torystructurewasmaintainedandthecomputationloadwassavedbyabout50％ThecomputationresultwithMatlabindicatesthatthemodifiedmodelandmethodarefeasible，andofpracticalvaluefortheanalysisofmetabolicactivityKEYWORDS：reactionnetwor