SVR-PAIRWISE method to predict MHC-II binding peptides

Abstract

Peptides binding to MHC molecules is of great importance in the process of triggering and initiating immune responses. There are two main kinds of MHC molecules, MHC class I and class II, and the prediction of MHC-II binding peptides is much more difficult due to their variable lengths, which makes it difficult to construct a preferable prediction model by using most of the existing methods. This paper presents a method, called as SVR-PAIRWISE, to combine Support Vector Regression (SVR) and pairwise alignment, to quantitatively predict the MHC-II binding peptides. The comparison results with some popular methods show its satisfying performances.