Employing Protein Language Models to Predict TCR-Peptide-MHC Binding: A Comprehensive Methods Guide.
A new methods chapter offers a practical guide to TCR–peptide–MHC binding prediction — the computational layer that determines whether a predicted neoantigen is actually recognized by T cells. It covers protein language model-era tools including TCR-ESM, NetTCR-2.2, ERGO-II, pMTnet, and ATM-TCR, plus ESM-2-, tcrLM-, and graph neural network-based approaches, with protocols for data preparation, execution, interpretation, and tool selection by research objective. For vaccine developers, this is a benchmarking reference for the immunogenicity-prediction step downstream of every neoantigen selection pipeline.