Rapid whole-genome sequencing for surveillance of Salmonella enterica serovar enteritidis

Abstract

For Salmonella enterica serovar Enteritidis, 85% of isolates can be classified into 5 pulsed-field gel electrophoresis (PFGE) types. However, PFGE has limited discriminatory power for outbreak detection. Although whole-genome sequencing has been found to improve discrimination of outbreak clusters, whether this procedure can be used in real-time in a public health laboratory is not known. Therefore, we conducted a retrospective and prospective analysis. The retrospective study investigated isolates from 1 confirmed outbreak. Additional cases could be attributed to the outbreak strain on the basis of whole-genome data. The prospective study included 58 isolates obtained in 2012, including isolates from 1 epidemiologically defined outbreak. Whole-genome sequencing identified additional isolates that could be attributed to the outbreak, but which differed from the outbreak-associated PFGE type. Additional putative outbreak clusters were detected in the retrospective and prospective analyses. This study demonstrates the practicality of implementing this approach for outbreak surveillance in a state public health laboratory.

Keywords: Salmonella enterica serovar Enteritidis; bacteria; high-throughput nucleotide sequencing; infectious disease outbreaks; public health laboratory surveillance; pulsed-field gel electrophoresis; whole-genome sequencing.

Figure 1

Maximum-likelihood tree of population structure of Salmonella enterica serovar Enteritidis isolates obtained in New York and neighboring states, USA. The tree was inferred by using a general time-reversible model with a gamma distribution and was inferred to be the best fit model by the maximum-likelihood method implemented in MEGA 5.1 (22). Values on branches are bootstrap values based on 150 bootstrap replicates. Note the well supported and distant cluster associated with the long-term care facility (LTCF), as well as additional clusters A and B. Labels of isolates are colored according to their New York State Department of Health Wadsworth Laboratories multilocus variable-number tandem-repeat analysis (MLVA) subtype designation. Green, MLVA subtype B; red, MLVA subtype W; black, MLVA subtype AE. + indicates isolates in the LTCF cluster that were detected only by whole-genome analysis and were not detected epidemiologically. Scale bar indicates single-nucleotide polymorphisms per site.

Figure 2

Maximum-likelihood tree of population structure of Salmonella enterica serovar Enteritidis isolates obtained in New York and neighboring states, USA. The tree was inferred by using a general time-reversible model with a gamma distribution, which was inferred to be the best fit model by the maximum-likelihood method implemented in MEGA 5.1 (22). Values on branches are bootstrap values based on 150 bootstrap replicates. Pulsed-field gel electrophoresis (PFGE) types are indicated on branches. Labels of isolates are colored according to their New York State Department of Health Wadsworth Laboratories multilocus variable-number tandem-repeat subtype designation. Green, MLVA subtype B; red, MLVA subtype W; orange, MLVA subtype AE; blue, MLVA subtype CR; black, MLVA subtype data missing and isolates from Allard et al. (11). Rectangles indicate well-supported clusters of at least 3 isolates, letters within the rectangles correspond to the cluster designation in the Table. LTCF, long-term care facility. Scale bar indicates single-nucleotide polymorphisms per site.