SISTR

SISTR performs in silico typing and serovar prediction of Salmonella genomes using core-genome multilocus sequence typing (cgMLST) and O- and H-antigen gene detection from whole-genome sequencing data.

Key Features:

• cgMLST-based typing: Uses core-genome multilocus sequence typing as a primary approach for sequence typing and clustering of Salmonella isolates.
• O- and H-antigen gene detection: Searches for O and H antigen genes, including wzx and wzy, to support serovar prediction.
• Sequence type and cluster methods: Integrates sequence type and cluster-based methods to assign serovars and group related isolates.
• Whole-genome sequencing input: Operates on whole-genome sequencing data for computational serotyping.
• Comparative performance: Demonstrated correct typing of 94% of 1,624 isolates spanning 72 serovars in a retrospective comparison against SeqSero, SeqSero2, and Metric Oriented Sequence Typer (MOST).
• Robustness to GC bias: cgMLST approach provides resilience against GC-biased sequencing data.
• Monophasic-variant limitation: Mapping-based, allele-mode tools such as SeqSero and SeqSero2 were found more reliable for predicting monophasic variants compared to SISTR's methods.
• Library-preparation impact: O-antigen detection can be affected by library preparation due to low GC content in genes such as wzx and wzy.
• Operational characteristics: Enables rapid and cost-efficient computational serotyping suitable for high-throughput genomic workflows.

Scientific Applications:

• Public health surveillance: Automated serovar prediction for routine Salmonella surveillance and monitoring.
• Outbreak investigation: Typing and clustering of isolates to support detection and investigation of Salmonella outbreaks.
• Food and animal source monitoring: Typing diverse isolates derived from animal and foodstuff sources to track transmission and contamination.
• Method validation and inter-laboratory comparison: Comparative evaluations to assess concordance with traditional serology and other in silico tools for data compatibility across laboratories.

Methodology:

SISTR employs core-genome multilocus sequence typing (cgMLST), searches for O- and H-antigen genes (including wzx and wzy), and uses sequence type and cluster-based methods; it has been compared against mapping-based, allele-mode tools such as SeqSero and SeqSero2.