%0 Journal Article
%T On Predicting Conformational B-cell Epitopes: a Comparative Study and a New Model
%J American Journal of Bioinformatics Research
%@ 2167-6976
%D 2011
%I 
%R 10.5923/j.bioinformatics.20110101.02
%X Identification of conformational B-cell epitopes is considered the crucial step in designing effective peptide vaccines against pathogens. Computer based methods play an important role in this process as the actual experimental determination of epitopes is very expensive in terms of cost and time. In this paper, we have carried out a comparative study and discussions for different methods based on the two major computational approaches for predicting conformational B-cell epitopes: sequence- based and structure- based approaches. As a result of this study, we developed a novel computational method ¡°CBCPRED¡± to predict conformational B-cell epitope residues from the target antigen structure by combining support vector machine model with protein structural features and the propensity scores of amino acid physico ¨C chemical properties. Using fivefold cross validation and leave-one-out cross validation techniques on the 75 antigen structures of the Discotope dataset, CBCPRED achieves an area under receiver operator characteristics curve (AUC) of 0.818 and 0.859, respectively. We benchmark ¡°CBCPRED¡± on a more recent benchmark (Ponomarenko et al. 2007) dataset after removing antigens sequence redundancy where no two antigen sequences have more than 40% sequence identity, achieving AUC of 0.747. CBCPRED is available at http://www.fci.cu.edu.eg:8080/CBCPRED/predict.html.
%K Conformational B-Cell Epitopes
%K Temperature Factor
%K Relative Solvent Accessibility
%K Propensity Score
%K Amino Acid Physic ¨C Chemical Properties
%K PSSM Profiles
%K Graph Centrality
%K Support Vector Machine Model
%U http://article.sapub.org/10.5923.j.bioinformatics.20110101.02.html