Salmonella spp. in silico serotyping with short reads and long reads

Salmonella infections are one of the leading causes for food poisoning. The identification of serovars of Salmonella achieved by their diverse surface proteins (Lipopolysaccharide (LPS) and protein) can reveal important information about their pathogenicity. Whole genome sequencing (WGS) and in silico serotyping is proposed as an alternative method to molecular methods for the surveillance of Salmonella outbreaks. So far, WGS data generated with Illumina sequencing is used to validate in silico methods. Oxford Nanopore Technologies (ONT) sequencing becomes more common and opens the ability to sequence ultra-long reads. For this study, we take ONT WGS data of Salmonella to investigate the performance of the tools SISTR and SeqSero2 compared to Illumina WGS data. The ONT data achieved the highest accuracy for both tools with 94% and 92% for SISTR and SeqSero2 respectively. Therefore, ONT WGS data can be an effective method for Salmonella serotyping and surveillance.