P2-138 Foodborne Outbreak Detection: Florida Department of Agriculture and Consumer Services' WGS SNP Pipeline in Action

Tuesday, July 11, 2017
Exhibit Hall (Tampa Convention Center)
S. Brian Caudle , Florida Department of Agriculture and Consumer Services , Tallahassee , FL
Carl Franconi, Jr. , Florida Department of Agriculture and Consumer Services , Tallahassee , FL
Serena Giovinazzi , Florida Department of Agriculture and Consumer Services , Tallahassee , FL
Amy Bryant , Florida Department of Agriculture and Consumer Services , Tallahassee , FL
Jason Crowe , Florida Department of Agriculture and Consumer Services , Tallahassee , FL
Introduction: Whole Genome Sequencing (WGS), together with its applications, is revolutionizing food safety. WGS provides a depth, specificity, and sensitivity previously unseen in outbreak detection technology by allowing isolate relationships to be elucidated by evolutionary descent. Due to the computational complexity required for WGS analysis, data interpretation is an obstacle for state labs, who are reliant on federal analysis pipelines.

Purpose: FDACS Bureau of Food Laboratories (BFL) has applied current technologies to develop an internal genomics pipeline to pinpoint possible foodborne outbreaks. This study demonstrates BFL’s ability to compare genomic data from clinical, food, and environmental Listeria monocytogenes isolates to deduce SNP-based relationships.

Methods: FL DOH genomic data from eight clinical L. monocytogenes isolates was compared against BFL’s L. monocytogenes phylogenic tree, containing approximately 700 Florida food and environmental isolates. Internal isolates were sequenced on a MiSeq using the Nextera XT and MiSeq V2 500-Cycle chemistry kits. An in-house developed pipeline consisting of kSNP3, FastTree2, and FigTree was used to establish preliminary sequence comparisons. The internally applied CFSAN SNP pipeline was used to detect hqSNPs amongst the cluster of interest. Lastly, FastTree2 was reincorporated along with FigTree to identify SNP distances and produce a phylogenetic tree.

Results: A cluster containing food and environmental isolates was found to be the closest to seven of the eight clinical samples. However, hqSNP analysis for this group showed that there were between 6,400 and 6,500 SNPs separating the clinical and food and environmental isolates. There were 1 to 4 SNPs identified within the cluster of seven clinical isolates.

Significance: BFL’s usage of an in-house SNP pipeline quickly disproved a possible correlation eight clinical L. monocytogenes isolates had to current food and environmental isolates within Florida. This highlights the BFL’s capability to promptly identify in-state foodborne outbreaks using a high resolution hqSNP-based approach.