P035 A method for large-scale analysis of HLA genetic variation

Abstract

Aim Less than 7% of HLA genes in IMGT/HLA are reported with full-length sequences. Recent high-throughput advances with NGS offers efficient full-length HLA typing. ∼160,000 HLA loci were reported by HML. A fully automatic annotation pipeline is needed to predict HLA exon–intron architecture and characterize genetic variation within exons and introns. Methods A program hml2fa was created to parse consensus sequences from HML to fasta format, then Clustal-Omega was applied for multiple sequence alignment with references. ExonFinder was created to predict exon–intron structures. UniqueSeq was used to remove sequence duplications for uniqueness analysis. Meanwhile, alignment sequence weights were calculated then bar plotted for validation. Results Sequence weight distribution indicated good alignment coverage especially in exon regions. Uniqueness analysis revealed thousands of full-length HLA variants that are potential novel alleles. In consideration of unique CDS and protein products, ∼50% of the variation was found in exons, resulting in 20–25% non-synonymous changes in class I HLA genes. Further analyses showed that exons 2 and 3 have significant high variability (p Conclusions This annotation pipeline can be used to predict HLA exon–intron structure and explore genetic variation on a large-scale. Surprisingly large numbers of intronic variants may affect HLA function and impact selection of matches for best transplant outcomes. Download : Download high-res image (307KB) Download : Download full-size image