Predicting binding affinity using differential evolution

During the processing and presentation of the exogenous antigen the binding between antigens and major histocompatibility complex class II molecules (MHC-II) is the precondition for activating helper T cell to response. For one kind of given MHC-II molecules, if the binding peptide can be predicted precisely, not only it can help us to understand the basic principle of the immune but also is useful and significant for the development of vaccine and the cure of autoimmune diseases. In this paper a method which is composed by Differential Evolution (DE) and position-specific scoring matrix (PSSM) is used to train and analyze the 14 kinds of MHC-II allele´s data and set up the prediction models separately. Comparing with other models the model called DE-PSSM in this paper performs better.