Challenges Associated with Investigating Salmonella Enteritidis with Low Genomic Diversity in New York State: The Impact of Adjusting Analytical Methods and Correlation with Epidemiological Data

Abstract

Defining investigation-worthy genomic clusters among strains of Salmonella Enteritidis is challenging because of their highly clonal nature. We investigated a cluster identified by core genome multilocus sequence typing (cgMLST) consisting of 265 isolates with isolation dates spanning two and a half years. This cluster experienced chaining, growing to a range of 14 alleles. The volume of isolates and broad allele range of this cluster made it difficult to ascertain whether it represented a common-source outbreak. We explored laboratory-based methods to subdivide and refine this cluster. These methods included using cgMLST with a narrower allele range, whole genome multilocus sequence typing (wgMLST) and high-quality single-nucleotide polymorphism (hqSNP) analysis. At each analysis level, epidemiologists retroactively reviewed exposures, geography, and temporality for potential commonalities. Lowering the threshold to 0 alleles using cgMLST proved an effective method to refine this analysis, resulting in this large cluster being subdivided into 34 smaller clusters. Additional analysis by wgMLST and hqSNP provided enhanced cluster resolution, with the majority of clusters being further refined. These analysis methods combined with more stringent allele thresholds and layering of epidemiologic data proved useful in helping to subdivide this large cluster into actionable subclusters.

Keywords: PulseNet; Salmonella Enteritidis; genomic epidemiology; whole-genome sequencing.

FIG. 1.

Analysis methods used to refine persisting Salmonella Enteritidis cluster in New York State. The initial cgMLST cluster of Salmonella Enteritidis was defined as 3 isolates within 5 alleles within 60 days. Retrospectively, this cluster was then refined into cgMLST subclusters by lowering the allele threshold to 0 alleles within 60 days. Further refinement of the 0 allele cgMLST subclusters containing 3 or more isolates was performed by wgMLST analysis of those subclusters with a threshold of 0 wgMLST alleles. The wgMLST subclusters were additionally assessed by hqSNP analysis. cgMLST, core genome multilocus sequence typing; hqSNP, high-quality single-nucleotide polymorphism; wgMLST, whole-genome multilocus sequence typing.

FIG. 2.

The maximum likelihood tree for the 265 SALM1.0 − 6743.2.4 × samples. The cgMLST- and wgMLST-clustered samples investigated are color coded and indicated by the text in the leaves. The tree is rooted at the midpoint and SRR19148189 is an outlier sample (not part of the SALM1.0 − 6743.2.4 × ). Excluding this outlier sample the entire SNP diversity of the tree is 33 SNPs. Branches with greater than 80% support are indicated by a star. Asterisk = reference genome.