No Assembly Required: Using BTyper3 to Assess the Congruency of a Proposed Taxonomic Framework for the Bacillus cereus Group With Historical Typing Methods

Abstract

The <i>Bacillus cereus</i> group, also known as <i>B. cereus sensu lato</i> (<i>s.l.</i>), is a species complex comprising numerous closely related lineages, which vary in their ability to cause illness in humans and animals. The classification of <i>B. cereus s.l.</i> isolates into species-level taxonomic units is essential for facilitating communication between and among microbiologists, clinicians, public health officials, and industry professionals, but is not always straightforward. A recently proposed genomospecies-subspecies-biovar taxonomic framework aims to provide a standardized nomenclature for this species complex but relies heavily on whole-genome sequencing (WGS). It thus is unclear whether popular, low-cost typing methods (e.g., single- and multi-locus sequence typing) remain congruent with the proposed taxonomy. Here, we characterize 2,231 <i>B. cereus s.l.</i> genomes using a combination of <i>in silico</i> (i) average-nucleotide identity (ANI)-based genomospecies assignment, (ii) ANI-based subspecies assignment, (iii) seven-gene multi-locus sequence typing (MLST), (iv) single-locus <i>panC</i> group assignment, (v) <i>rpoB</i> allelic typing, and (vi) virulence factor detection. We show that sequence types (STs) assigned using MLST can be used for genomospecies assignment, and we provide a comprehensive list of ST/genomospecies associations. For <i>panC</i> group assignment, we show that an adjusted, eight-group framework is largely, albeit not perfectly, congruent with the proposed eight-genomospecies taxonomy, as <i>panC</i> alone may not distinguish (i) <i>B. luti</i> from Group II <i>B. mosaicus</i> and (ii) <i>B. paramycoides</i> from Group VI <i>B. mycoides</i>. We additionally provide a list of loci that capture the topology of the whole-genome <i>B. cereus s.l.</i> phylogeny that may be used in future sequence typing efforts. For researchers with access to WGS, MLST, and/or <i>panC</i> data, we showcase how our recently released software, BTyper3 (https://github.com/lmc297/BTyper3), can be used to assign <i>B. cereus s.l.</i> isolates to taxonomic units within this proposed framework with little-to-no user intervention or domain-specific knowledge of <i>B. cereus s.l.</i> taxonomy. We additionally outline a novel method for assigning <i>B. cereus s.l.</i> genomes to pseudo-gene flow units within proposed genomospecies. The results presented here highlight the backward-compatibility and accessibility of the recently proposed genomospecies-subspecies-biovar taxonomic framework and illustrate that WGS is not a necessity for microbiologists who want to use the proposed nomenclature effectively.