基于自相关函数的蛋白质-蛋白质作用位点预测研究

Predicting protein-protein interaction site based on auto-correlation function

蛋白质-蛋白质相互在细胞生命过程扮演重要角色,广泛参与免疫反应,信号传导,基因表达,蛋白质合成等,研究蛋白质-蛋白质作用位点,将有助于揭示生命过程的许多本质,对预防、诊断疾病,以及突变设计、蛋白质相互作用网络构建等方面均具有重要的参考价值。根据蛋白质-蛋白质作用位点残基倾向性及作用位点与其周围临近残基密切相关的特性,本文提出基于序列谱(或空间谱)构建自相关函数,度量邻近残基之间的相关程度,采用AdaBoost分类器预测蛋白质-蛋白质相互作用位点,精度达到67.6%,表明本文的方法预测蛋白质-蛋白质相互作用位点是有效的,为研究蛋白质-蛋白质相互作用位点研究提供了一种新方法。

Protein-protein interactions play an important role in many cellular processes such as immune regulation,signal transduction,gene translation,and protein synthesis.The study of protein-protein interaction site can help us to reveal many essential problems in the life processes,and it also plays a key role in the disease prevention,disease diagnosis,mutant design and protein-protein interaction network construction,etc.According to the protein-protein interface propensity and the correlations among the adjacent residues,a novel method for prediction the protein-protein interaction site was proposed,which used the auto-correlation function to measure the correlation of adjacent residues from the sequence profiles（or spatial profiles）.The overall success rate arrived at 67.6%with AdaBoost algorithm.The simulation results show that this method can predict the protein-protein interaction site effectively.