Deep learning algorithms have been utilized to achieve enhanced performance in pattern-recognition tasks. The ability to learn complex patterns in data has tremendous implications in immunogenomics. T-cell receptor (TCR) sequencing assesses the diversity of the adaptive immune system and allows for modeling its sequence determinants of antigenicity. We present DeepTCR, a suite of unsupervised and supervised deep learning methods able to model highly complex TCR sequencing data by learning a joint representation of a TCR by its CDR3 sequences and V/D/J gene usage. We demonstrate the utility of deep learning to provide an improved ‘featurization’ of the TCR across multiple human and murine datasets, including improved classification of antigen-specific TCRs and extraction of antigen-specific TCRs from noisy single-cell RNA-Seq and T-cell culture-based assays. Our results highlight the flexibility and capacity for deep neural networks to extract meaningful information from complex immunogenomic data for both descriptive and predictive purposes.

SARS-CoV-2 (CoV2) antibody therapies, including COVID-19 convalescent plasma (CCP), monoclonal antibodies, and hyperimmune globulin, are among the leading treatments for individuals with early COVID-19 infection. The functionality of convalescent plasma varies greatly, but the association of antibody epitope specificities with plasma functionality remains uncharacterized. We assessed antibody functionality and reactivities to peptides across the CoV2 and the 4 endemic human coronavirus (HCoV) genomes in 126 CCP donations. We found strong correlation between plasma functionality and polyclonal antibody targeting of CoV2 spike protein peptides. Antibody reactivity to many HCoV spike peptides also displayed strong correlation with plasma functionality, including pan-coronavirus cross-reactive epitopes located in a conserved region of the fusion peptide. After accounting for antibody cross-reactivity, we identified an association between greater alphacoronavirus NL63 antibody responses and development of highly neutralizing antibodies against CoV2. We also found that plasma preferentially reactive to the CoV2 spike receptor binding domain (RBD), versus the betacoronavirus HKU1 RBD, had higher neutralizing titer. Finally, we developed a 2-peptide serosignature that identifies plasma donations with high anti-spike titer, but that suffer from low neutralizing activity. These results suggest that analysis of coronavirus antibody fine specificities may be useful for selecting desired therapeutics and understanding the complex immune responses elicited by CoV2 infection.

Allergic reactions occur when IgE molecules become crosslinked by antigens such as food proteins. Here we create the ‘AllerScan’ programmable phage display system to characterize the binding specificities of anti-allergen IgG and IgE antibodies in serum against thousands of allergenic proteins from hundreds of organisms at peptide resolution. Using AllerScan, we identify robust anti-wheat IgE reactivities in wheat allergic individuals but not in wheat-sensitized individuals. Meanwhile, a key wheat epitope in alpha purothionin elicits dominant IgE responses among allergic patients, and frequent IgG responses among sensitized and non-allergic patients. A double-blind, placebo-controlled trial shows that alpha purothionin reactivity, among others, is strongly modulated by oral immunotherapy in tolerized individuals. AllerScan may thus serve as a high-throughput platform for unbiased analysis of anti-allergen antibody specificities.

Antigen discovery technologies have largely focused on major histocompatibility complex (MHC) class I-restricted human T cell receptors (TCRs), leaving methods for MHC class II-restricted and mouse TCR reactivities relatively undeveloped. Here we present TCR mapping of antigenic peptides (TCR-MAP), an antigen discovery method that uses a synthetic TCR-stimulated circuit in immortalized T cells to activate sortase-mediated tagging of engineered antigen-presenting cells (APCs) expressing processed peptides on MHCs. Live, tagged APCs can be directly purified for deconvolution by sequencing, enabling TCRs with unknown specificity to be queried against barcoded peptide libraries in a pooled screening context. TCR-MAP accurately captures self-reactivities or viral reactivities with high throughput and sensitivity for both MHC class I-restricted and class II-restricted TCRs. We elucidate problematic cross-reactivities of clinical TCRs targeting the cancer/testis melanoma-associated antigen A3 and discover targets of myocarditis-inciting autoreactive T cells in mice. TCR-MAP has the potential to accelerate T cell antigen discovery efforts in the context of cancer, infectious disease and autoimmunity.