DocMeetings

Clinical outcomes are correlated with aggregate B (BCR) and T cell receptor (TCR) diversity (the adaptome) in several infectious diseases and cancers.  Advances in dimer avoidance multiplexed PCR (DAM-PCR) followed by next-generation sequencing (NGS) enable measurements of immune repertoire diversity and clonality, allowing prediction of cancer states and response to treatment. However, current diversity measurements generate an overwhelming amount of information (up to 1025 possible CDR3 variable region sequences ), which can mask cancer-specific and predictive information contained within small 3-6- amino-acid long motifs..  While deep learning algorithms are capable of powerful data-driven predictions, these approaches typically require large patient cohorts to make accurate predictions. Here we present a natural language processing model (NLP) model for early prediction of patient response to IL-2 immunotherapy capable of modeling a small patient cohort based on motif elucidation of CDR3 TCR and BCR clonotypes. Furthermore, we demonstrate the utility of analyzing all seven TCR and BCR chains for early detection and monitoring of patient states, providing new mechanistic insight into TCR and BCR orchestration during treatment.