Machine Learning Analysis of Naïve B-Cell Receptor Repertoires Stratifies Celiac Disease Patients and Controls

Abstract

Celiac disease (CeD) is a common autoimmune disorder caused by an abnormal immune response to dietary gluten proteins. The disease has high heritability. HLA is the major susceptibility factor, and the HLA effect is mediated via presentation of deamidated gluten peptides by disease-associated HLA-DQ variants to CD4+ T cells. In addition to gluten-specific CD4+ T cells the patients have antibodies to transglutaminase 2 (autoantigen) and deamidated gluten peptides. These disease-specific antibodies recognize defined epitopes and they display common usage of specific heavy and light chains across patients. Interactions between T cells and B cells are likely central in the pathogenesis, but how the repertoires of naïve T and B cells relate to the pathogenic effector cells is unexplored. To this end, we applied machine learning classification models to naïve B cell receptor (BCR) repertoires from CeD patients and healthy controls. Strikingly, we obtained a promising classification performance with an F1 score of 85%. Clusters of heavy and light chain sequences were inferred and used as features for the model, and signatures associated with the disease were then characterized. These signatures included amino acid (AA) 3-mers with distinct bio-physiochemical characteristics and enriched V and J genes. We found that CeD-associated clusters can be identified and that common motifs can be characterized from naïve BCR repertoires. The results may indicate a genetic influence by BCR encoding genes in CeD. Analysis of naïve BCRs as presented here may become an important part of assessing the risk of individuals to develop CeD. Our model demonstrates the potential of using BCR repertoires and in particular, naïve BCR repertoires, as disease susceptibility markers.

Different B cell subpopulations show distinct patterns in their IgH repertoire metrics

“…Recent IgH repertoire studies have moved towards using machine learning and artificial intelligence in contrast to traditional statistical approaches for goals including vaccine design, immunodiagnostics and antibody discovery ( Greiff et al, 2017 ; Ostmeyer et al, 2017 ; Konishi et al, 2019 ; Shemesh et al, 2021 ). Previous work has focused on representing repertoires as sequence or subsequence-based features, that is, overlapping amino acid k-mers and their Atchley biophysiochemical properties ( Greiff et al, 2017 ; Ostmeyer et al, 2017 ).…”

Individualized VDJ recombination predisposes the available Ig sequence space

“…Machine learning is increasingly used for AIRR classification both on the sequence (Akbar et al 2021;Friedensohn et al 2020;Isacchini et al 2021;Greiff et al 2017b) and repertoire-level (Shemesh et al 2021;Emerson et al 2017;Sidhom et al 2021;Pavlović et al 2021). Future studies will need to investigate whether differences in RGM also impact repertoire classification (Greiff et al 2020).…”

“…In the future, it will be interesting to study whether individual differences in RGMP lead to differences in the propensity to generate antigen-specific (for instance auto-reactive) sequences (Shemesh et al 2021) and, consequently, to the existence of individualized holes in the repertoire (Perelson and Oster 1979). These analyses will require large-scale naïve (unselected) and disease-linked AIRR-seq data (Watson, Glanville, and Marasco 2017).…”

Immune2vec: Embedding B/T Cell Receptor Sequences in ℝN Using Natural Language Processing

“…CDR3 sequences were embedded into R 100 , yielding a matrix of size [number of sequences × 100] where each row r i represents a sequence i and each column C j represents the value of the embedding in dimension j(j ∈ {1…100}). To characterize whole repertoires using CDR3 embedded vectors we applied an analogous clustering approach to the one used in ( 11) and (24). Vectors were first grouped according to their V-gene, J-gene, and CDR3 length.…”