Versiti Blood Center of Wisconsin Milwaukee, Wisconsin, United States
Background/Case Studies: The application of nanopore long read DNA sequencing for clinical diagnostics was evaluated, focused on HLA genotyping as a first venture into the technology. Short read next-generation sequencing (Illumina MiSeq) is highly accurate for HLA genotyping but requires complex sample preparation and long sequencing time, and the short reads can result in ambiguous cis/trans phasing of HLA genotypes. This study assessed the concordance of nanopore HLA genotyping data compared to short read HLA MiSeq. Key nanopore quality control metrics were also evaluated in addition to the ability of HLA nanopore technology to resolve HLA ambiguities.
Study
Design/Methods: Genotypes for 11 HLA loci (HLA-A, -B, -C, -DRB1, DRB3/4/5, -DQA1, -DQB1, -DPA1, -DPB1) were generated with 2 different HLA nanopore reagent/software configurations, using DNAs previously sequenced via MiSeq. A set of 48 [n=40 from whole blood (WB); n=8 from buccal swabs] was tested with reagents and analysis software from Omixon (Budapest, HU)in conjunction with compatible library prep reagents, flow cell, MinION Mk1B device and MinKNOW data acquisition software from Oxford Nanopore Technologies (Oxford, UK). Nanopore HLA genotypes generated were compared to HLA MiSeq reference genotypes. To further explore nanopore HLA sequencing options, a subset n=10 of the WB-derived DNAs was retested in a 2nd study phase using reagents and analysis software from GenDx (Utrecht, NL) in conjunction with compatible high-accuracy Oxford reagents and flow cell, and device and acquisition software as described above. The 10 samples were tested in singlicate, HLA genotypes determined and compared to reference MiSeq genotypes. Additional analyses were performed for the 10 samples only, examining the ability of nanopore long reads to resolve ambiguous MiSeq HLA genotypes, along with compiling key MinKNOW performance metrics N50 and sequencing time.
Results/Findings: For both nanopore HLA sequencing pipelines, 100% concordance was observed (2-field resolution) at all 11 HLA loci and for all n=48 samples when results were compared to the MiSeq HLA reference data. No allelic dropouts were observed. For the follow-up 10 samples only, which underwent additional data analysis, short sequencing times (mean 0.71hr) were observed to reach the data endpoint of 20k reads per singlicate run. Additionally for the 10 samples, significant reductions in ambiguities were observed for most HLA loci due to the high N50 (mean 4.61kb) and improved phasing of distant polymorphisms with nanopore technology relative to the HLA MiSeq data.
Conclusions: In this feasibility study, nanopore high-resolution HLA sequencing demonstrated 100% concordance when HLA genotyping results were compared to predicate MiSeq HLA genotypes, and improvements were observed for sequencing time, N50, and ambiguity resolution. Nanopore-based HLA sequencing is a promising solution for rapid and accurate HLA genotyping.
