Accurate and efficient whole-genome sequencing (WGS) is crucial for clinical diagnostics and surveillance of bacterial infections. Here, we investigate the potential of a new Oxford Nanopore Technologies (ONT) workflow for WGS of clinically relevant bacterial isolates. Specifically, we assess the performance of R10.4.1 flow cells in combination with the V14 version of the transposase-based (RBK) library preparation kit to provide rapid and accurate genomic epidemiological comparisons of bacterial species of clinical importance. We focused on retrospective collections of outbreak-associated Corynebacterium diphtheriae (CDIP) and vancomycin-resistant enterococci (VRE) and benchmarked expected performance parameters such as genome assembly quality, genotyping [multi-locus sequence typing (MLST) and core-genome multi-locus sequence typing (cgMLST)], SNP profiling and antimicrobial resistance and virulence prediction, against WGS data obtained routinely by Illumina MiSeq sequencing. Complete concordance with Illumina results was observed for MLST in both species and for cgMLST in CDIP, across all ONT kits and software evaluated. For VRE, however, cgMLST results varied with strain identity, library preparation kit and analysis parameters, likely due to software challenges to correctly call methylated bases. Yet, the use of the latest basecalling models combined with a PCR-based library preparation kit (RPB) reliably reproduced Illumina cgMLST results across all tested VRE strains. By testing two hybrid strategies combining PCR-free and PCR-based library preparation approaches, we also showed that combining PCR-free and PCR-based methods may yield a promising strategy, achieving both high accuracy and assembly completeness. Genomic-based antimicrobial resistance (AMR) prediction was consistent across sequencing methods, and we further highlight advantages and limitations of the PCR-based, PCR-free and mixed assemblies, to inform on the genomic context of AMR genes. This study demonstrates that a Nanopore-only sequencing approach may offer improved accuracy and consistency for classical bacterial typing in outbreak investigations, paving the way to wider use in clinical microbiology laboratories.
