Gibbs motif sampler, weight matrix and artificial neural network for the prediction of MHC class-II binding peptides

Abstract

The identification of MHC class-II restricted epitope is an important goal in peptide based vaccine and diagnostic development. In the present study, we discuss the applications of Gibbs motif sampler, weight matrix and artificial neural network for the prediction of peptide binding to sixteen MHC class-II molecules of human and mouse. The average prediction performances of sixteen MHC class-II molecules in terms of Aroc, based on Gibbs motif sampler, sequence weighting and artificial neural network are 0.56, 0.55 and 0.51 respectively. However, further improvements in the performance of software tools for prediction of MHC class-II binding peptide based on various methods largely depends on the size of training and validation datasets and the correct identification of the peptide binding core. © 2009 Springer Berlin Heidelberg.