Fig. 2.

Similarity in the CGR Community. Similarity between the 1310 members of the CGR community, calculated using the number of shared hashes in their Mash sketches. In comparison to the ZYMO + MBARC-26 community (Additional file 1: Fig. 1), there are many clusters of high similarity in the CGR community. Genomes within these clusters are very difficult to distinguish between, and contribute to the lower classification accuracy, across all classification approaches, in this community