Title: Trans-allelic model for prediction of peptide:MHC-II interactions

Abstract: Major histocompatibility complex class two (MHC-II) molecules are trans-membrane proteins and key components of the cellular immune system. Upon recognition of foreign peptides expressed on the MHC-II binding groove, helper T cells mount an immune response against invading pathogens. Therefore, mechanistic identification and knowledge of physico-chemical features that govern interactions between peptides and MHC-II molecules is useful for the design of effective epitope-based vaccines, as well as for understanding of immune responses. In this paper, we present a comprehensive trans-allelic prediction model, a generalized version of our previous biophysical model, that can predict peptide interactions for all three human MHC-II loci (HLA-DR, HLA-DP and HLA-DQ), using both peptide sequence data and structural information of MHC-II molecules. The advantage of this approach over other machine learning models is that it offers a simple and plausible physical explanation for peptide-MHC-II interactions. We train the model using a benchmark experimental dataset, and measure its predictive performance using novel data. Despite its relative simplicity, we find that the model has comparable performance to the state-of-the-art method. Focusing on the physical bases of peptide-MHC binding, we find support for previous theoretical predictions about the contributions of certain binding pockets to the binding energy. Additionally, we find that binding pockets P 4 and P 5 of HLA-DP, which were not previously considered as primary anchors, do make strong contributions to the binding energy. Together, the results indicate that our model can serve as a useful complement to alternative approaches to predicting peptide-MHC interactions.