基于残基序列信息的蛋白质相互作用位点预测(英文)  

Predicting protein interaction sites with residue sequence information

作　　者：彭新俊[1] 王翼飞[2] 

机构地区：[1]上海师范大学计算数学系,上海200234 [2]上海大学数学系,上海200444

出　　处：《计算机与应用化学》2008年第6期649-653,共5页Computers and Applied Chemistry

基　　金：Supported by the National Natural Science Foundation of China(30571059);the National High-Tech Research and Develop-ment Program of China(2006AA022190).

摘　　要：蛋白质相互作用位点在细胞进程中有着非常重要的作用.尽管利用高通量方法发现蛋白质相互作用位点取得很大的成功,仍需要计算方法辅助预测实验中的相互作用位点.本文提出了基于残基序列谱、进化率和疏水性的预测异源蛋白质复合物作用位点的两种向量表示方法并以支持向量机实现预测.其中,提出新的向量表示法取得更好的预测性能.文中的数据集由66个异源复合物蛋白质链组成.Protein-protein interaction sites play a crucial role in the cellular process. Despite advances in high-throughput methods for discovering proteinprotein interaction sites, the continued computational methods to help direct experimentalists in the search were need for predicting interaction sites. This paper proposed two vector representation methods for predicting protein interaction sites in heterocomplexes using residue sequence profile, evolutionary rate, and hydrophobicity in support vector machine （SVM）, which are the novel vector representation and the common vector representation, respectively. Compared the two SVM predictors using different vector representation methods, the prediction performance of the former predictor is obviously higher than that of the common method. The study was based on a non-redundant data set of heterodimers consisting of 66 protein chains.

关 键 词：支持向量机 蛋白质相互作用位点 序列谱 进化率 疏水性 向量表示 

分 类 号：Q811.4[生物学—生物工程] O235[理学—运筹学与控制论]