Hello,
I’m working on oral Streptococcus species using the GTDB species clusters names and have had some discussions about how the 95% ANI cut-off splits some of these species into a very large number of species clusters, which can be argued to artificially split well-delineated species. Two examples of this are S. mitis and S. oralis, which have extended out into double-letter species designations (ie, S. oralis_BW). I suspect Strep are particularly tricky to define clear boundaries for because of their extensive recombination, but also the current ways to taxonomically classify an isolate or MAG are not always reliable - I’ve identified several genomes in NCBI labeled S. mitis that should be S. oralis, and vice versa, based on phylogenetic analyses.
The general agreement amongst people I’ve worked with on this is that the ANI boundaries for Streptococcus species clusters (and some other oral genera) should be reduced to around 93%. This regroups many of the smaller species clusters into larger species clusters that more closely resemble currently accepted streptococcal taxonomic species. I’ve read in the GTDB publications about slightly raising the ANI cut-off where appropriate, but didn’t see anything about lowering it. I was wondering if there are discussions about applying lower ANI cut-offs in the GTDB for particular species or within specific genera (specifically Streptococcus, but also Campylobacter and possibly Actinomyces)?
Thanks!