Figure 4. Species annotation vs groups.

The 15 groups produced by cutting the sketching and FastANI hierarchies (outer circle) coincide well with species annotation (inner circle), fixing the few apparently mislabeled genomes. The numbers to the right indicate number of groups sharing the species name. Three groups remained without a species name. To improve the illustration, the confirmed K. pneumoniae dataset was reduced by keeping representatives from clusters obtained at a Mash distance of 0.005.