HLA‐C genetic diversity and evolutionary insights in two samples from Brazil and Benin

Abstract

Human leukocyte antigen‐C (HLA‐C) is a classical HLA class I molecule that binds and presents peptides to cytotoxic T lymphocytes in the cell surface. HLA‐C has a dual function because it also interacts with Killer‐cell immunoglobulin‐like receptors (KIR) receptors expressed in natural killer and T cells, modulating their activity. The structure and diversity of the HLA‐C regulatory regions, as well as the relationship among variants along the HLA‐C locus, are poorly addressed, and few population‐based studies explored the HLA‐C variability in the entire gene in different population samples. Here we present a molecular and bioinformatics method to evaluate the entire HLA‐C diversity, including regulatory sequences. Then, we applied this method to survey the HLA‐C diversity in two population samples with different demographic histories, one highly admixed from Brazil with major European contribution, and one from Benin with major African contribution. The HLA‐C promoter and 3′UTR were very polymorphic with the presence of few, but highly divergent haplotypes. These segments also present conserved sequences that are shared among different primate species. Nucleotide diversity was higher in other segments rather than exons 2 and 3, particularly around exon 5 and the second half of the 3′UTR region. We detected evidence of balancing selection on the entire HLA‐C locus and positive selection in the HLA‐C leader peptide, for both populations. HLA‐C motifs previously associated with KIR interaction and expression regulation are similar between both populations. Each allele group is associated with specific regulatory sequences, reflecting the high linkage disequilibrium along the entire HLA‐C locus in both populations.

Whole-genome sequencing of 1,171 elderly admixed individuals from the largest Latin American metropolis (São Paulo, Brazil)

“…For the 1000 Genomes dataset, we obtained high coverage BAM files using the ASPERA protocol. We processed these BAM files using hla-mapper version 4 17 (www.castelli-lab.net/apps/hla-mapper), as described elsewhere 18,19 .…”

“…We processed each HLA locus separately. For variant refinement and selection, we used the vcfx checkpl, checkad, and evidence algorithms to introduce missing alleles in genotypes with low likelihood and annotate each variant with a series of quantitative parameters 19 (www.castelli-lab.net/apps/vcfx). Each variant that has not been approved by the vcfx evidence algorithm was evaluated manually.…”

“…We previously developed hla-mapper 37 to optimize mappings for HLA genes, providing high-confidence genotype and haplotype calls for this unusually polymorphic region 38 , with complex structure involving duplications. We applied hla-mapper in the SABE dataset, detecting 2.4X more variants in the HLA class I genes than with the computational workflow for genotype calling used in the entire genome.…”

KIR2DL4 genetic diversity in a Brazilian population sample: implications for transcription regulation and protein diversity in samples with different ancestry backgrounds

“…However, the end products of these alignments were quite unsatisfactory, with large numbers of erroneously mapped reads due to the high sequence similarity among KIR genes [6]. Therefore, we adopted a strategy similar to the previously used for HLA genes on this same sample to obtain optimized mappings [43][44][45]. We used hla-mapper version 3.07 (available at www.castelli-lab.net/apps/hla-mapper) but with a specific database for KIR genes [43].…”

“…We used GATK HaplotypeCaller [47,48] (version 4.1.7) to call genotypes in the GVCF mode, concatenating all genotypes in a single VCF file using GATK GenotypeGVCFs. Variant refinement was performed using vcfx checkpl and vcfx evidence as discussed elsewhere [44]. This step ensures that only high-quality genotypes are moved forward to the haplotyping step.…”