The Bayesian optimist's guide to adaptive immune receptor repertoire analysis

Abstract

Probabilistic modeling is fundamental to the statistical analysis of complex data. In addition to forming a coherent description of the data-generating process, probabilistic models enable parameter inference about given datasets. This procedure is well developed in the Bayesian perspective, in which one infers probability distributions describing to what extent various possible parameters agree with the data. In this paper, we motivate and review probabilistic modeling for adaptive immune receptor repertoire data then describe progress and prospects for future work, from germline haplotyping to adaptive immune system deployment across tissues. The relevant quantities in immune sequence analysis include not only continuous parameters such as gene use frequency but also discrete objects such as B-cell clusters and lineages. Throughout this review, we unravel the many opportunities for probabilistic modeling in adaptive immune receptor analysis, including settings for which the Bayesian approach holds substantial promise (especially if one is optimistic about new computational methods). From our perspective, the greatest prospects for progress in probabilistic modeling for repertoires concern ancestral sequence estimation for B-cell receptor lineages, including uncertainty from germline genotype, rearrangement, and lineage development.

Functional Enrichment and Analysis of Antigen-Specific Memory B Cell Antibody Repertoires in PBMCs

“…During the acute phase of pathogen exposure, naïve B cells enter germinal centers and undergo immunoglobulin (Ig) receptor somatic hypermutation (SHM) to increase antigen binding, isotype class switching to IgG or IgA for effector activities, and differentiation to short-lived antibody secreting plasmablasts or to long-lived plasma and memory B cells (2). Successful diversification of B cell clones and their corresponding Ig receptors creates clonal families, each a cluster or lineage of related antibodies all descended from the naïve B cell ancestor (3), and possessing variations in antigen binding and functional activities. Antibodies secreted from plasma cells, which mainly reside in the bone marrow, provide steady state protection against repeat infections.…”

The expanding role of systems immunology in decoding the T cell receptor repertoire

“…While such tools have so far facilitated substantial progress in our understanding of TCR repertoires, there is a growing need for the development of more sophisticated computational approaches to provide higher-dimensional insights into TCR biology and to ensure statistically robust interpretation of data observations. Regarding the latter, there has been recent discussion to motivate the use of probabilistic models that provide an ensemble view of immune repertoires and opportunities to capture the stochastic nature and complexity of the TCR repertoire system [31,32]. It is also expected that future attentions will become more focused towards the development of computational and statistical methods that are more closely integrated with the experimental design of TCR sequencing strategies to facilitate specific interrogation of TCR repertoires, as well as experimental approaches that target specific mechanisms impacting TCR repertoires.…”

Weakly supervised identification and generation of adaptive immune receptor sequences associated with immune disease status

“…Choi 2022, Ruffolo et al 2021, Olsen et al 2022b trained larger generative models on individual antibody sequences. Marcou et al 2018 used an explicit Bayesian probabilistic model (Olson and Matsen 2018), to generate AIR repertoires that mimic experimental ones regardless of any sequence labels, while Davidsen et al 2019 and Isacchini et al 2021 used deep generative models for the same goal. Most importantly, Pradier et al 2023 designed a generative model AIRIVA that can learn a low-dimensional interpretable representation of TCR repertoires with respect to the repertoire labels.…”