Chapter 4 - Predicting protein interaction sites from unlabeled sample information based on a semi-supervised approach

Abstract

There is only a small part of protein interactions had been identified by experiments, which will cause the lack of prediction accuracy and generalization ability of predictors in protein interaction sites prediction. This chapter introduces three semi-supervised support vector machine–based methods to improve the performance in the protein interaction sites prediction, in which the information of unlabeled protein sites can be involved. Herein, five features related with the evolutionary conservation of amino acids are extracted from HSSP database and Consurf Sever to represent the residues within the protein sequence. The experimental results demonstrated that unlabeled information can effectively improve prediction performance of protein interaction sites and the best prediction performance, that is, the accuracy of 70.7%, the sensitivity of 62.67%, and the specificity of 78.72%, respectively. With comparison to the existing studies, the semi-supervised models show the improvement of the predication performance.