A framework for annotation of antigen specificities in high-throughput T-cell repertoire sequencing studies

Abstract

Recently developed molecular methods allow large-scale profiling of T-cell receptor (TCR) sequences that encode for antigen specificity and immunological memory of these cells. However, it is well known, that the even unperturbed TCR repertoire structure is extremely complex due to the high diversity of TCR rearrangements and multiple biases imprinted by VDJ rearrangement process. The latter gives rise to the phenomenon of "public" TCR clonotypes that can be shared across multiple individuals and non-trivial structure of the TCR similarity network. Here we outline a framework for TCR sequencing data analysis that can control for these biases in order to infer TCRs that are involved in response to antigens of interest. Using an example dataset of donors with known HLA haplotype and CMV status we demonstrate that by applying HLA restriction rules and matching against a database of TCRs with known antigen specificity it is possible to robustly detect motifs of an epitope-specific responses in individual repertoires. We also highlight potential shortcomings of TCR clustering methods and demonstrate that highly expanded TCRs should be individually assessed to get the full picture of antigen-specific response.Recent developments in the field of bioinformatic analysis of AIRR-Seq data are aimed at providing a mean for annotation of TCR repertoires with predicted antigen specificities. For example, lists pathogen-and disease-associated TCRs and VDJdb database [12] features a large set of TCRs with experimentally verified epitope specificities and their MHC restrictions. Existing computational methods for TCR repertoire annotation allow both matching against a database of known antigen specificities [12], [13] and clustering of TCR sequences for de novo motif detection [4], [14]. Annotation of a large number of TCR repertoires from healthy donors [15], [16] demonstrates both high variance of frequencies of epitope-specific T-cells and the imprint of past and ongoing pathogen encounters. Thus, de novo discovery of Tcells associated with antigens of interest or certain disease appears to be a hard problem, complicated by the biases in the structure of the naive (unperturbed) TCR repertoire [17], presence of existing clonal expansions specific to unrelated pathogens and high number of false positives that result from the extremely high diversity of the TCR repertoire.