Assessing a single targeted next generation sequencing for human leukocyte antigen typing protocol for interoperability, as performed by users with variable experience

Abstract Background A simplified protocol for HLA-typing -by NGS, developed for use with the Illumina MiSeq, was performed by technologists with varying NGS experience to assess accuracy and reproducibility. Methods Technologists from six laboratories typed the same 16 samples at HLA-A, B, C, DRB1, and DQB1. The protocol includes long range PCR, library preparation, and paired-end 250 bp sequencing. Two indexing strategies were employed: locus-specific indexing whereby each locus was tagged uniquely and sample-specific indexing whereby all 5 loci for a sample were pooled prior to library preparation. Sequence analysis was performed with two software packages, Target HLA (Omixon) and NGSengine (GenDx). Results The average number of sequence reads per library was 387,813; however, analysis was limited to 40,000 reads for locus-indexed libraries and 200,000 reads for sample-indexed libraries resulting in an average depth of coverage of 1444 reads per locus. Sufficient reads for genotype analysis were obtained for 98.4% of libraries. Genotype accuracy was >97% in pooled amplicons and >99% in individual amplicons by both software analysis. Inter-laboratory reproducibility was 99.7% and only cause of discordance was cross-contamination of a single amplicon. Conclusions This NGS HLA-typing protocol is simple, reproducible, scalable, highly accurate and amenable to clinical testing.