BcePred: prediction of continuous B-cell epitopes in antigenic sequences using physico-chemical properties

Abstract

A crucial step in designing of peptide vaccines involves the identification of B-cell epitopes. In past, numerous methods have been developed for predicting continuous B-cell epitopes, most of these methods are based on physico-chemical properties of amino acids. Presently, its difficult to say which residue property or method is better than the others because there is no independent evaluation or benchmarking of existing methods. In this study the performance of various residue properties commonly used in B-cell epitope prediction has been evaluated on a clean dataset. The dataset used in this study consists of 1029 non-redundant B cell epitopes obtained from Bcipep database and equally number of non-epitopes obtained randomly from SWISS-PROT database. The performance of each residue property used in existing methods has been computed at various thresholds on above dataset. The accuracy of prediction based on properties varies between 52.92% and 57.53%. We have also evaluated the combination of two or more properties as combination of parameters enhance the accuracy of prediction. Based on our analysis we have developed a method for predicting B cell epitopes, which combines four residue properties. The accuracy of this method is 58.70%, which is slightly better than any single residue property. A web server has been developed to predict B cell epitopes in an antigen sequence. The server is accessible from http://www.imtech.res.in/raghava/bcepred/.