Development of a wgMLST method for investigating foodborne botulism cases through genome comparisons

Abstract

Clostridium botulinum , which can cause foodborne botulism through production of botulinum toxin in contaminated foods, is classified as a Tier 1 select agent in the United States. Since the ability to share isolates is federally regulated, approaches to sharing molecular data, including whole genome sequence data (WGS), are needed. We have implemented a whole genome multilocus sequence typing (wgMLST) database for surveillance of Group I C. botulinum (serotypes A, B, and F) using WGS to compare strains on a national level during botulism investigations or events. Five-hundred forty-two isolates associated with botulism cases were sequenced and included in the database, including technical replicates. WGS were generated using Illumina (2 x 250 bp MiSeq and 2 x 150 bp MiniSeq) and Ion Torrent (1 x 200 bp PGM and 1 x 400 bp S5) DNA sequencing. Raw read quality, assembly quality, and allele calling metrics were evaluated. wgMLST was compared with High quality SNP-typing (hqSNPs), a currently accepted method, on a subset of five botulism clusters or events. Illumina produced the best quality sequence data ( Q scores ≥ 30) and the highest number of allele calls. IonTorrent data had quality issues that impacted allele calls and clustering of epidemiologically related isolates. Therefore, sequencing using the Illumina platform will be supported moving forward as part of our national workflow. wgMLST and hqSNPs produced comparable results using Illumina data, and related strains could be distinguished from unrelated strains that were genetically similar and harbored the same toxin subtype. Therefore, a central, wgMLST database will allow U.S. public health laboratories to analyze their own sequencing data and submit data to a national database in PulseNet, the U.S. National molecular surveillance system for foodborne, waterborne, and One Health related outbreaks or events.