多特征融合的蛋白质相互作用位点预测

Prediction of protein interaction sites using multi-feature amalgamation

蛋白质相互作用位点预测为蛋白质功能和药物设计的理解提供重要线索。而蛋白质的各种特征为蛋白质相互作用位点预测提供了大量有用信息,特别是进化信息、残基序列邻近和空间邻近性。不同的蛋白质特征对蛋白质间的相互作用的贡献也不一样。通过提取蛋白质序列谱、保守性和残基熵,提出了特征融合技术对蛋白质相互作用位点进行研究,采用SVM构建三种预测器,分别对各种不同的特征加以验证,实验结果表明了基于特征融合方法的有效性和正确性。

Prediction of protein-protein interaction sites provides the key dues to understand the function of a protein and drug design.However,many different features of protein provide much information for the prediction of protein-protein interaction sites, especially protein-protein evolution information,residues sequence neighbor list and residues space neighbor list.Different features are not the same for prediction of protein-protein interaction sites.Choosing protein profile of sequence alignment,residues conserved score and residue entropy,suggesting the technology of protein features amalgamation to predict protein-protein interaction sites.This paper adopts SVM to construct three predictors which validate them with different features as input vector, the experiments show the validation and correctness of the method based on feature amalgamation.