Introduction: Tumor-immune dynamics are often interrogated using hybridized capture-based sequencing approaches which require the use of baits that target the protein-coding regions of the genome. However, the genes encoding the antigen presentation machinery, including human leukocyte antigen (HLA) proteins, are very polymorphic, resulting in thousands of unique genotypes across patient cohorts. Here, using whole exome sequencing (WES) and hybridized capture-based RNA sequencing (RNAseq) from melanoma patients, we show that these genes are not uniformly assessed by common sequencing methods, and require careful analysis for accurate detection of copy number and somatic alterations.

Methods: WES and RNAseq data was analyzed from 760 melanoma tumor biopsies, with WES derived from patient-matched normal blood samples. Coverage metrics for 26 antigen presentation genes, including classical and non-classical HLA genes, were evaluated, and the difference in coverage was calculated for heterozygous alleles. Each individual was assessed for heterozygous single nucleotide polymorphisms (SNPs) within each allele, and with respect to the haplotype of the human reference genome.

Results: Focusing on the 26 genes assessed, HLA-A, HLA-B, HLA-C, and HLA-DRB1 were the most diverse across our patient cohort; there was an average of 134 unique pairs of HLA alleles in these genes. We compared the SNP profiles of allele combinations to the haplotype of the human reference genome, and observed four patterns: heterozygous alleles are similar to each other and the reference haplotype, heterozygous alleles are different from each other and different to the reference, alleles are very similar to each other and greatly differ from the reference, or one allele is more similar to the reference haplotype. The fourth pattern showed the greatest bias in sequencing, where the allele more similar to the human reference genome had a median 4-fold increase in coverage in normal WES data.

Conclusions: Capture-based sequencing shows selection bias, as exome probes have been designed based upon the reported human reference genome, but results in selection bias in hybridized capture sequencing approaches. Alleles that have higher polymorphic differences with respect to the reference are not equally captured, resulting in nonuniform sequencing of heterozygous alleles.

Citation Format: Egmidio Medina, Katie Campbell, Antoni Ribas. Hybridized-capture sequencing methods bias analysis of antigen presentation genes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 765.