BSFINDER: Finding Binding Sites of HCV Proteins Using a Support Vector Machine

Abstract

Hepatitis C virus (HCV) infection is a major cause of liver disease and a dangerous threat to public health. Hence, the problem of finding interactions between HCV and human proteins has received much attention. In this paper, we present an approach to predicting binding residues in HCV proteins using a support vector machine (SVM) classifier. Based on six biochemical properties of amino acids (sequence profile, accessible surface area, residue binding propensity, sequence entropy, hydrophobicity and conservation weight), the SVM classifier achieved an average accuracy of 93%. Contiguous residues in the sequence act together to determine a binding site, and a window of 11 residues (the target residue and 5 adjacent residues on each side) gave the best result in our study. Our approach has been implemented in a program called BSFinder (Binding Site Finder), which is available at http://wilab.inha.ac.kr/bsfinder. BSFinder will be of considerable help in predicting binding residues and potential interacting partners of a protein.