Individualized VDJ recombination predisposes the available Ig sequence space

Abstract

The process of recombination between variable (V), diversity (D), and joining (J) immunoglobulin (Ig) gene segments determines an individual's naive Ig repertoire and, consequently, (auto)antigen recognition. VDJ recombination follows probabilistic rules that can be modeled statistically. So far, it remains unknown whether VDJ recombination rules differ between individuals. If these rules differed, identical (auto)antigen-specific Ig sequences would be generated with individual-specific probabilities, signifying that the available Ig sequence space is individual specific. We devised a sensitivity-tested distance measure that enables inter-individual comparison of VDJ recombination models. We discovered, accounting for several sources of noise as well as allelic variation in Ig sequencing data, that not only unrelated individuals but also human monozygotic twins and even inbred mice possess statistically distinguishable immunoglobulin recombination models. This suggests that, in addition to genetic, there is also nongenetic modulation of VDJ recombination. We demonstrate that population-wide individualized VDJ recombination can result in orders of magnitude of difference in the probability to generate (auto)antigen-specific Ig sequences. Our findings have implications for immune receptor–based individualized medicine approaches relevant to vaccination, infection, and autoimmunity.

Genetic variation in the immunoglobulin heavy chain locus shapes the human antibody repertoire

“…This study provides support for the idea that leveraging IG genetic data can better delineate Ab response dynamics in a variety of contexts. For one, there is growing interest in developing predictive models for V(D)J recombination and repertoire diversity 84, 85 , and applying Ab repertoire profiling as a diagnostic tool for disease and clinical phenotypes of high public health relevance 86, 87 . However, current models do not explicitly account for genetic factors, and the effects of this on model performance are not known 84, 85 .…”

“…For one, there is growing interest in developing predictive models for V(D)J recombination and repertoire diversity 84, 85 , and applying Ab repertoire profiling as a diagnostic tool for disease and clinical phenotypes of high public health relevance 86, 87 . However, current models do not explicitly account for genetic factors, and the effects of this on model performance are not known 84, 85 . Our results indicate that future work in this area should explore ways to integrate genetic data; this will likely be critical for better understanding commonalities and differences in repertoire signatures, not only for gene usage patterns, but also in identifying additional features (e.g., public clonotypes 1,2 ), overall leading to improved metrics for immune response monitoring and prediction modeling.…”

“…In addition, as cohorts increase in size, additional insight will come from the consideration of other variables such as genetic ancestry, positive/negative selection, age, B cell subset and tissue 97–99 . Finally, the models utilized here could be extended to assess the contribution of IGH polymorphisms to other repertoire signatures, including N/P addition and CDR3 features, which also are influenced by heritable factors 20, 21, 38, 85 .…”

Machine Learning Approaches to TCR Repertoire Analysis

“…In addition, some HLA variants are associated with onset of autoimmune diseases (18). These results contrast with those of immunoglobulin for which V(D)J recombination process before selection is found to be highly different even between monozygotic (MZ) twins (19). Moreover, as HLA genes are highly variant (12), TCRs that bind to the same peptide can differ between people.…”

Comparing T cell receptor repertoires using optimal transport

“…Most current methods of repertoire comparison involve reducing the TCR sequences into simpler summaries and then comparing these summaries [2][3][4][5][6][7][8][9][10]. As comparing full CDR3 sequences can be highly involved, one approach is to simply compare CDR3 length distributions [11,12].…”