Abstract
BACKGROUND: In the United States, most Clostridium botulinum type A strains isolated during laboratory investigations of human botulism demonstrate the presence of an expressed type A botulinum neurotoxin (BoNT/A) gene and an unexpressed BoNT/B gene. These strains are designated type A(B). The most common pulsed-field gel electrophoresis (PFGE) pattern in the C. botulinum PulseNet database is composed of A(B) strains. The purpose of this study was to evaluate the ability of genome sequencing and multi-loci variable number of tandem repeat analysis (MLVA) to differentiate such strains. RESULTS: The genome sequences of type A(B) strains evaluated in this study are closely related and cluster together compared to other available C. botulinum Group I genomes. In silico multilocus sequence typing (MLST) analysis (7-loci) was unable to differentiate any of the type A(B) strains isolated from seven different outbreak investigations evaluated in this study. A 15-locus MLVA scheme demonstrated an improved ability to differentiate these strains, however, repeat unit variation among the strains was restricted to only two loci. Reference-free single nucleotide polymorphism (SNP) analysis demonstrated the ability to differentiate strains from all of the outbreaks examined and a non-outbreak associated strain. CONCLUSIONS: This study confirms that type A(B) strains that share the same PFGE pattern also share closely-related genome sequences. The lack of a complete type A(B) strain representative genome sequence hinders the ability to assemble genomes by reference mapping and analysis of SNPs at pre-identified sites. However, compared to other methods evaluated in this study, a reference-free SNP analysis demonstrated optimal subtyping utility for type A(B) strains using de novo assembled genome sequences.