Major Histocompatibility Complex (MHC) molecules play important roles in adaptive immune response. An important function of MHC molecules is to bind peptide fragments derived from pathogens and to display them on the cell surface for being recognized by the appropriate T cells [1]. This stimulates subsequent immune response in order to fight against these pathogens. The MHC gene family is mainly divided into two subgroups: class I (MHC-I) and class II (MHC-II). Compared to predicting MHC-I binding peptides, predicting MHC-II binding peptides is more challenging, since the binding grooves are open at the both ends.

Computational methods are used to quickly select a small number of promising peptides for further biochemical experimental verification. They can be divided into two groups: allele-specific and pan-specific methods [2]. In the allele-specific methods, both training and test peptides are from the same MHC molecule. In contrast, pan-specific methods can predict the binding peptides of MHC molecules that have very few or even no training data. According to underlying techniques used, the allele-specific methods can be roughly divided into four groups: position specific score matrix (PSSM) based methods, artificial neural network (ANN) based methods, kernel based methods, and multiple instance learning based methods.

We develop a novel multiple instance learning based method, called MHC2MIL, for MHC-II peptide binding prediction by considering peptide flanking region and residue positions. MHC2MIL is an allele specific method. Now we provide 26 alleles for prediction.


  1. Janeway J CA, Travers P, Walport M, et al Immunobiology: The Immune System in Health and Disease. Garland Science Publishing, New York., 5 edition 2001.
  2. Zhang L, Udaka K, Mamitsuka H, Zhu S. 2012 Toward more accurate pan-specific MHC-peptide binding prediction: a review of current methods and tools.. Brief Bioinform. 13(3):350-64.