Study On The Analysis And Identification Of The Disordered Regions Of Intrinsically Disordered Proteins

In the natural state,the intrinsically disordered protein is a kind of protein that has no stable three-dimensional structure,but still can perform normal biological functions.Many proteins with this property have been found in the experiment,which are different from the traditional protein paradigm,so it is called natural disordered protein.But the intrinsically disordered protein mainly plays the function in the disordered regions,these regions distribute many function sites.Therefore,predicting and analyzing of the disordered regions of intrinsically disordered proteins have become a hot issue.According to DisProt database,we divided intrinsically disordered proteins into six functions,in which included chaperones,molecular assembly,molecular recognition scavengers,entropic chain,modification site,molecular recognition effectors.We choose the disordered sequences of six molecular functions as the research object and established the six sub-databases.Then,we extracted four feature parameters separately in the six sub-databases,namely,the amino acid index(AAIndex),codon,protein secondary structure(PSS)and chemical shifts(ACS),and adopted these features to predict and analyze the disordered regions based on the molecular functions by support vector machine(SVM).The results by jackknife test were as follows:1)In the above four single feature information,the chemical shifts(ACS)was optimal feature and the overall accuracy was 99.29%;2)Among the feature fusion,the overall accuracy can reach 39.92%and 42.94%by using AAIndex+Codon and AAIndex+3PSS as features.And the overall accuracy was up to 41.73%by using AAIndex+Codon+3PSS as features.Compared with the prediction of the single feature,the accuracy of was improved after the feature fusion.In addition,the disordered proteins were classified into six classes based on the molecular functions of intrinsically disordered proteins(IPDs)in DisProt database.Five residues were selected nearby the ordered and disordered junction site respectively for the six kinds of functional proteins and the differences of 20 amino acids and 6 kinds of hydrophilic and hydrophobic amino acids were analyzed by the variance analysis.The results showed that there existed significant differences in amino acid content and hydrophilic and hydrophobic characteristic in the junction regions of the different functional proteins.The amino acids(D,G,H,I,P,S,T and Y)had different distributions in the ordered regions while the amino acids(D,G and V)had different distributions in the disordered regions.Strongly hydrophobic amino acids(L,I,V,A,M and F),weakly hydrophilic or weakly hydrophobic amino acids(S,T,Y and W)as well as amino acids P and G had different distributions in the ordered regions while the amino acid G had different distributions in the disordered regions.