Laboratoire Matière et Systèmes Complexes

Modelling and predicting antigen presentation with Restricted Boltzmann Machines

Barbara Bravi

Abstract: Immune recognition of infected and malignant cells requires ’’presentation’’ on their surface of antigens (i.e. short peptides) by human leukocyte antigen class I (HLA-I) proteins, which are coded by one of the most polymorphic alleles in the human genome. At the molecular level, antigen presentation consists of a protein-protein binding process between antigens and HLA-I proteins. The identification of clinically relevant neoantigens (antigens harbouring tumour-specific mutations) is currently a highly sought-after goal in designing novel cancer immunotherapeutic strategies and algorithms aimed at predicting peptide binding to HLA-I proteins are therefore valuable tools to accelerate the validation of putative neoantigens. To tackle this problem, we resort to a statistical mechanics framework of inference from aminoacid sequences based on Restricted Boltzmann Machines, probabilistic graphical models characterized by a layer of ``feature’’ variables. We show that this framework provides an efficient, interpretable, generative modelling approach to antigen presentation that allows one to accurately infer from aminoacid sequence what antigens can bind to HLA-I proteins along with their HLA-type specificity.

Tuesday February 18 at 12:00 in room 646A