Predicting the antigenic peptidome

Antigenic peptides are short protein sequences that are recognized by the immune system. The Major Histocompatibility Complex (MHC) mediates this process. MHC class I molecules bind intracellular peptides and transport them to the cell surface, where peptide-specific interactions with killer (CD8+) T-cells are induced. Peptide presentation involves a complex transport pathway characterized by checkpoints and proofreading mechanisms. Here, we show that cell metabolism can strongly affect antigen presentation. We develop a mechanistic model that predicts presentation probabilities from peptide-specific data - expression level and MHC binding affinity - and the efficiency of the transport pathway depending, in turn, on the metabolic state of the cell. Together, these factors determine the quantity and quality of the presented peptidome; that is, the total amount and the specific distribution of peptides available for recognition by T cells. We validate the model by combining genome-wide expression data and mass-spectrometry data of presented peptides across different cell lines. We discuss characteristic shifts of the presented peptidome between cancer cell lines and infer underlying cell state changes, including variations in peptide transport and MHC expression.