TABLE 3.
Revised accuracy of different methods on GEBA, FDA-ARGROS, and NCTC 3000 datasets.
| Expecteda | 16Sb | KmerFinder (Com)c | KmerFinder (type)d | fIDBACe | fIDBAC-ANI (95%)f | |
| GEBA | 973 | 617 (63.41%) | 272 (27.95%) | 908 (93.32%) | 963 (98.97%) | 963 (98.97%) |
| FDA-ARGROS | 505 | 364 (72.08%) | 478 (94.65%) | 429 (84.95%) | 501 (99.21%) | 488 (96.63%) |
| NCTC-3000 | 967 | 644 (66.60%) | 880 (91.00%) | 671 (69.38%) | 962 (99.48%) | 951 (98.34%) |
| Total | 2,445 | 1,625 (66.46%) | 1,630 (66.67%) | 2,008 (82.12%) | 2,427 (99.26%) | 2,402 (98.24%) |
aExpected, the number of query genomes.
b16S, number of 16S top 1 result that matches the original label.
cKmerFinder (Com), number of KmerFinder top 1 result with its complete genome database that matches the original label.
dKmerFinder (type), number of KmerFinder top 1 result with its type-strain genome database that matches the original label.
efIDBAC, number of fIDBAC top 1 result with its curated genome database that matches the original label.
ffIDBAC, number of fIDBAC top 1 result with its curated genome database that matches the 95% ANI threshold and the original label.