TCRen: predicting TCR recognition of unseen epitopes based on residue-level pairwise statistical potential

Prediction of TCR-peptide interactions has great importance for therapy of cancer, infectious and autoimmune diseases, but remains a major challenge, particularly for unseen epitopes. We present a structure-based method that enables scoring of TCR-peptide interactions using an energy potential (TCRen) derived from statistics of TCR-peptide contacts in existing crystal structures. We show that TCRen has high performance in discriminating cognate/unrelated peptides and can facilitate the identification of cancer neoepitopes recognized by tumor-infiltrating lymphocytes.