Subtyping evaluation of Salmonella Enteritidis using SNP and core genome MLST with nanopore reads

Whole genome sequencing (WGS) for public health surveillance and epidemiological investigation of foodborne pathogens predominantly relies on sequencing platforms that generate short reads. Continuous improvement of long-read nanopore sequencing such as Oxford Nanopore Technologies (ONT) presents a potential for leveraging multiple advantages of the technology in public health and food industry settings, including rapid turnaround and onsite applicability in addition to superior read length. However, evaluation, standardization and implementation of the ONT approach to WGS-based, strain-level subtyping is challenging, in part due to its relatively high base-calling error rates and frequent iterations of sequencing chemistry and bioinformatic analytics. Using an established cohort of Salmonella Enteritidis isolates for subtyping evaluation, we assessed the technical readiness of ONT for single nucleotide polymorphism (SNP) analysis and core-genome multilocus sequence typing (cgMLST) of a major foodborne pathogen. By multiplexing three isolates per flow cell, we generated sufficient sequencing depths under seven hours of sequencing for robust subtyping. SNP calls by ONT and Illumina reads were highly concordant despite homopolymer errors in ONT reads (R9.4.1 chemistry). In silico correction of such errors allowed accurate allelic calling for cgMLST and allelic difference measurements to facilitate heuristic detection of outbreak isolates. Our study established a baseline for the continuously evolving nanopore technology as a viable solution to high quality subtyping of Salmonella, delivering comparable subtyping performance when used standalone or together with short-read platforms.