短肽段構(gòu)象空間的降維與序列上不同位點之間殘基的相關(guān)性研究.pdf

1、中國科學(xué)技術(shù)大學(xué)碩士學(xué)位論文短肽段構(gòu)象空間的降維與序列上不同位點之間殘基的相關(guān)性研究姓名：盧文申請學(xué)位級別：碩士專業(yè)：計算生物學(xué)指導(dǎo)教師：劉海燕20060520AbstractThestructuresoffragmentsarethebaseofthestructureofproteinbecausethewholestructureoftheproteincanbeassembledfromthoseshortfragmentsWh

2、atweconcerniswhethertherealstructuremodelsonlyexistinsomepartoftheconformationalspaceInordertoconfirmit，inthechapter2，weusetheprincipalcomponentanalysistoconstructthenewcoordinatessystemInthenewcoordinatesspace，wecanuses

3、everalcoordinatestodescribetheprinciplevarietyoftherealconformationofsegmentsWealsoclusterthesegmentsinnewcoordinatespace，andanalysestheinformationofsequencesofeveryclusterAmpleevidencesuggeststhatthelocalstructuresofpep

4、tidefragmentsinnativeproteinsaretosomeextentencodedbytheirlocalsequencesDetectingsuchlocalcorrela—tionsisimportantbutitisstillanopenquestionwhatwouldbethemostappropriatemethodThisispartlybecausethatconventionalsequencean

5、alysestreataminoacidpreferencesateachsiteofaproteinsequenceindependentlywhileitisoftentheintersiteinteractionsthatbringaboutlocalsequence—structurecorrelationsInthechapter3，weintroduceanewschemetocapturethecorrelationbet

6、weenaminoacidpreferencesatdifferentsitesfordifferentlocalstructuretypesWeconstructalibraryofnineresiduefragments，dividethefragmentsintoclustersbasedontheirlocalstructuresForeachlocalstructureclusterortype，weusechisquaret

7、eststoidentifycorrelatedpreferencesofaminoacidcombinationsatpairsofsitesWeconstructascorefunctionincludingboththesinglesiteaminoacidpreferencesandthedualsiteaminoacidcombinationpreferences，whichcanbeusedtoidentifywhether

8、asequencefragmentwouldhaveastrongtendencytoformaparticularlocalstructureinnativeproteinsTheresultsshowedthat，first，givenalocalstructurepattem，dualsiteaminoacidcombinationscontainsdifferentinformationfromsinglesiteaminoac

9、idpreferences；second，wepresentrepresentativeexampleswhichshowthatmanyofthestatisticallyidentifiedcorrelationsagreewithpreviouslyproposedheuristicrulesaboutlocalsequencestructurecorrelations，orareconsistentwithphysical—ch