Whither Extensive Genomic-Based Microbial Taxonomic Revision?

In the study published by Potter et al. in this issue of Clinical Chemistry (1), 103 genomes under the auspices of either Gardnerella vaginalis or Gardnerella spp. contributed to the National Center for Biotechnology Information database were accessed with the goal of further elucidating subspecies or new species designations (termed genomospecies in the context of in silico analysis). This genus is one of several that can contribute to the dynamic disease entity of bacterial vaginosis. A Canadian group has recently identified differentiating genetic patterns between bacterial vaginosis-, aerobic vaginitis-, and Lactobacillus spp.-dominated vaginal microbiomes (2). Intricate genetic characterization of novel taxa within the Gardnerella genus may identify target organisms for subsequent studies of pathogenesis. In lieu of the traditional DNA-DNA hybridization standard, Potter et al. used 4 modalities of in silico analysis for subsequent genetic characterization of whole genome assemblies. These included 2 average nucleotide identity (ANI)3 platforms (3), tetranucleotide frequency (4), and average amino acid identity (AAI) (5). The authors used a conservative criterion for assignment of genomospecies (i.e., concordant results derived from ≥2 modalities of characterization). Classifications were supported by ancillary core genome, accessory genome, and metatranscriptome analyses. Using this algorithm, the authors report 9 Gardnerella genomospecies—verifying the taxonomic designation of Gardnerella piotii sp. nov. but suggesting potential conflicts in the taxonomic status of Gardnerella leopoldii sp. nov. and Gardnerella swidsinskii sp. nov. (6). Past literature has espoused prokaryotic species-level ANI cutoffs of 94% to 96% (3). In a recent publication, Ciufo et al. (7) evaluated this threshold by identifying 335 taxonomic designations for which at least 10 GenBank assemblies were available with ANI alignments above 10% coverage. These assemblies were compared with the submitted type taxon assembly and labeled as concordant (taxonomic agreement) or discordant (taxonomic disagreement). The average ANI for concordant comparisons was 97.1%, …