De novo Inference of Diversity Genes and Analysis of Non-canonical V(DD)J Recombination in Immunoglobulins

Abstract

The V(D)J recombination forms the immunoglobulin genes by joining the variable (V), diversity (D), and joining (J) germline genes. Since variations in germline genes have been linked to various diseases, personalized immunogenomics aims at finding alleles of germline genes across various patients. Although recent studies described algorithms for de novo inference of V and J genes from immunosequencing data, they stopped short of solving a more difficult problem of reconstructing D genes that form the highly divergent CDR3 regions and provide the most important contribution to the antigen binding. We present the IgScout algorithm for de novo D gene reconstruction and apply it to reveal new alleles of human D genes and previously unknown D genes in camel, an important model organism in immunology. We further analyze non-canonical V(DD)J recombination that results in unusually long CDR3s with tandem fused IGHD genes and thus expands the diversity of the antibody repertoires. We demonstrate that tandem CDR3s represent a consistent and functional feature of all analyzed immunosequencing datasets, reveal ultra-long CDR3s, and shed light on the mechanism responsible for their formation.

Automated analysis of immunosequencing datasets reveals novel immunoglobulin D genes across diverse species

“…Therefore, sometimes multiple extensions are desired from a single abundant k-mer. However, since it is not clear how to avoid false positives in the case of multiple extensions, the IgScout algorithm [28] uses long seed k-mers (that are unique among all D genes), thus bypassing the multiple extension problem. Although this approach works for species with partially known germline genes, it is unclear how to select k for species with unknown germline genes and short germline genes.…”

“…Since DiversityAnalyzer uses the set of known V and J genes to compute CDR3s and since V and J genes for camel and macaque are unknown, we used human V and J genes to construct CDR3s for these species. Since some CDR3s may be affected by sample preparation errors, we grouped CDR3s differing by at most 3 mismatches and constructed a consensus CDR3 for each group as described in [28]. Constructing consensus CDR3s also helps concentrate on only the recombinant diversity (and not SHMs) of immunosequencing datasets by removing CDR3s with SHMs to some extent.…”

“…The sequences encoded by D genes play important roles in B cell development, antigen binding site diversity, and antibody production [27]. Safonova and Pevzner [28] recently developed the IgScout algorithm for de novo inference of D genes using immunosequencing data. Unlike algorithms for de novo inference of V and J genes [23,24], it does not rely on alignments against closest germline genes that might lead to erroneous inferences [29,30].…”

immuneSIM: tunable multi-feature simulation of B- and T-cell receptor repertoires for immunoinformatics benchmarking

“…Among these sequences, 94% of the time the junction identified by immuneSIM was found to be identical to that of IMGT. The V and J annotation overlapped in >97% of simulated sequences, while D annotations, a generally more difficult problem due to deletions and insertions (Bolotin et al , 2015; Safonova and Pevzner, 2019), showed an overlap of ~60%. Taken together, these results support the notion that immuneSIM repertoires are nearly indistinguishable from experimental repertoires with respect to major statistical descriptors and thus can serve as a reliable basis for benchmarking immunoinformatics tools.…”

Systematic profiling of full-length immunoglobulin and T-cell receptor repertoire diversity in rhesus macaque through long read transcriptome sequencing

“…However, such protocols are optimized for applications in human and mouse [38] and have not yet been applied for Ig/TCR analysis in rhesus macaques. Moreover, even with MPCR or 5’ RACE it is still difficult to capture complete Ig mRNA transcripts with the commonly available sequencing instruments like the 2 x 250 bp HiSeq system and the 2 x 300 bp MiSeq system, in part due to the large length variability of recombined molecules [39] and the potential for non-canonical recombination events [40]. Library preparation techniques have recently been improved to support this capability but only in human [33].…”