A comparison of five Illumina, Ion Torrent, and nanopore sequencing technology-based approaches for whole genome sequencing of SARS-CoV-2
Carbo, Ellen C.,Mourik, Kees,Boers, Stefan A.,Munnink, Bas Oude,Nieuwenhuijse, David,Jonges, Marcel,Welkers, Matthijs R. A.,Matamoros, Sebastien,van Harinxma thoe Slooten, Joost,Kraakman, Margriet E. M.,Karelioti, Evita,van der Meer, David,Veldkamp, Karin Ellen,Kroes, Aloys C. M.,Sidorov, Igor,de Vries, Jutte J. C.
DOI: https://doi.org/10.1007/s10096-023-04590-0
2023-04-06
Abstract:Rapid identification of the rise and spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants of concern remains critical for monitoring of the efficacy of diagnostics, therapeutics, vaccines, and control strategies. A wide range of SARS-CoV-2 next-generation sequencing (NGS) methods have been developed over the last years, but cross-sequence technology benchmarking studies have been scarce. In the current study, 26 clinical samples were sequenced using five protocols: AmpliSeq SARS-CoV-2 (Illumina), EasySeq RC-PCR SARS-CoV-2 (Illumina/NimaGen), Ion AmpliSeq SARS-CoV-2 (Thermo Fisher), custom primer sets (Oxford Nanopore Technologies (ONT)), and capture probe-based viral metagenomics (Roche/Illumina). Studied parameters included genome coverage, depth of coverage, amplicon distribution, and variant calling. The median SARS-CoV-2 genome coverage of samples with cycle threshold (Ct) values of 30 and lower ranged from 81.6 to 99.8% for, respectively, the ONT protocol and Illumina AmpliSeq protocol. Correlation of coverage with PCR Ct values varied per protocol. Amplicon distribution signatures differed across the methods, with peak differences of up to 4 log 10 at disbalanced positions in samples with high viral loads (Ct values ≤ 23). Phylogenetic analyses of consensus sequences showed clustering independent of the workflow used. The proportion of SARS-CoV-2 reads in relation to background sequences, as a (cost-)efficiency metric, was the highest for the EasySeq protocol. The hands-on time was the lowest when using EasySeq and ONT protocols, with the latter additionally having the shortest sequence runtime. In conclusion, the studied protocols differed on a variety of the studied metrics. This study provides data that assist laboratories when selecting protocols for their specific setting.
infectious diseases,microbiology