Probing T-cell response by sequence-based probabilistic modeling

• Published in: PLoS computational biology Vol. 17; no. 9; p. e1009297
• Main Authors: Bravi, Barbara, Balachandran, Vinod P., Greenbaum, Benjamin D., Walczak, Aleksandra M., Mora, Thierry, Monasson, Rémi, Cocco, Simona
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 02.09.2021; PLOS; Public Library of Science (PLoS)
• Subjects: Amino acids; Antigen (tumor-associated); Antigens; Biology and Life Sciences; Cancer; Cancer Survivors; Carcinoma, Pancreatic Ductal - immunology; Cloning; Cluster Analysis; Computational Biology - methods; Cytokines; Datasets; Datasets as Topic; Epitopes; Humans; Immunotherapy; Infectious diseases; Information processing; Learning algorithms; Learning models (Stochastic processes); Life Sciences; Lymphocytes; Lymphocytes T; Machine learning; Mathematical models; Medicine and Health Sciences; Models, Statistical; Mutation; Neoantigens; Next-generation sequencing; Pancreatic Neoplasms - immunology; Patients; Peptides; Physiological aspects; Probabilistic models; Receptors, Antigen, T-Cell - immunology; Research and Analysis Methods; T cell receptors; T cells; T-cell receptor; T-Lymphocytes - immunology; Tumors; France
• ISSN: 1553-7358; 1553-734X; 1553-7358
• DOI: 10.1371/journal.pcbi.1009297

With the increasing ability to use high-throughput next-generation sequencing to quantify the diversity of the human T cell receptor (TCR) repertoire, the ability to use TCR sequences to infer antigen-specificity could greatly aid potential diagnostics and therapeutics. Here, we use a machine-learning approach known as Restricted Boltzmann Machine to develop a sequence-based inference approach to identify antigen-specific TCRs. Our approach combines probabilistic models of TCR sequences with clone abundance information to extract TCR sequence motifs central to an antigen-specific response. We use this model to identify patient personalized TCR motifs that respond to individual tumor and infectious disease antigens, and to accurately discriminate specific from non-specific responses. Furthermore, the hidden structure of the model results in an interpretable representation space where TCRs responding to the same antigen cluster, correctly discriminating the response of TCR to different viral epitopes. The model can be used to identify condition specific responding TCRs. We focus on the examples of TCRs reactive to candidate neoantigens and selected epitopes in experiments of stimulated TCR clone expansion.