Using average nucleotide identity to improve taxonomic assignments in prokaryotic genomes at the NCBI

Abstract

Average nucleotide identity analysis is a useful tool to verify taxonomic identities in prokaryotic genomes, for both complete and draft assemblies. Using optimum threshold ranges appropriate for different prokaryotic taxa, we have reviewed all prokaryotic genome assemblies in GenBank with regard to their taxonomic identity. We present the methods used to make such comparisons, the current status of GenBank verifications, and recent developments in confirming species assignments in new genome submissions.

Keywords: GenBank; RefSeq; taxonomy; type strains.

Fig. 1.

ANI process workflow for processing of pre-submission genomes.

Fig. 2.

Example of taxonomy correction markup on a GenBank record. This information was added after working closely with the submitter to correct the identification of the genome entry.

Fig. 3.

K-mer tree showing the distribution of Shigella genomes among those of E. coli. The bar indicates the percentage nucleotide rate over the length of the genome.

Fig. 4.

K-mer tree showing genome variability amongst Lactobacillus gasseri assemblies. The ANI cutoff of 93.5 % includes both groups of assemblies (a and b), whilst a 96 % cutoff will separate them. Type assemblies are highlighted. The bar indicates the percentage nucleotide substitution rate over the length of the genome.