Hypervariable T-cell receptors (TCR) play a key role in adaptive immunity, recognising a vast diversity of pathogen-derived antigens. High throughput sequencing of TCR repertoires (RepSeq) produces huge datasets of T-cell receptor sequences from blood and tissue samples. However, our ability to extract clinically relevant information from RepSeq data is limited, mainly because little is known about TCR-disease associations. Here we present a statistical approach called ALICE (Antigen-specific Lymphocyte Identification by Clustering of Expanded sequences) that identifies TCR sequences that are actively involved in the current immune response from a single RepSeq sample, and apply it to repertoires of patients with a variety of disorders - autoimmune disease (ankylosing spondylitis), patients under cancer immunotherapy, or subject to an acute infection (live yellow fever vaccine). The methods robustness is demonstrated by the agreement of its predictions with independent assays, and is supported by its ability to selectively detect responding TCR in the memory but not in the naïve subset. ALICE requires no longitudinal data collection nor large cohorts, and is thus directly applicable to most RepSeq datasets. Its results facilitate the identification of TCR variants associated with a wide variety of diseases and conditions, which can be used for diagnostics, rational vaccine design and evaluation of the adaptive immune system state.

Detecting T-cell receptors involved in immune responses from single repertoire snapshots