A User's Guide to the Bioinformatic Analysis of ANI and Other Genome Sequence Data for Prokaryotic Species Description

Abstract Recent large‐scale analysis of isolate genomes and metagenomes has revealed that prokaryotes form distinct clusters that can be equated to species. Members of the same species typically show >95–96% genome‐wide average nucleotide identity (ANI) to each other and <85% ANI to members of other species, while genomes with 85–96% ANI are much less frequent and likely represent organisms evolving toward a new species. That is, the ANI value distribution reveals a species‐level gap (or shortage of values) around 95–96% and below. This gap is the most common pattern observed among the best‐sampled 330 bacterial and 30 archaeal groups; however, exceptions to the pattern do exist, usually driven by specific underlying evolutionary and/or ecological processes. Therefore, the species ANI threshold should be adjusted if the ANI value distribution reveals a different range for the gap, as is sometimes the case for more clonal and/or recently evolved species. This article aims to provide a practical guide on how to evaluate the ANI value distribution to determine if a species gap exists for a group of genomes of interest and at what ANI values this gap manifests. The guide also includes discussion of the available approaches for estimating ANI as well as comparisons to alternative metrics for estimating whole‐genome relatedness. The article concludes with the existence of ANI gaps within these 360 species to define strains, genomovars, and phylogroups as well as a review of the literature recommendations for defining taxa higher than the species level and applications to other microbes and viruses.