TABLE 2.
Serovar prediction accuracy in E. coli serovars depends on the amount of representation in our reference databasea
| Correct | Incorrect | Indeterminate | Number in group | |
|---|---|---|---|---|
| 100–300 assemblies | 95.6% | 0.6% | 3.8% | 3 |
| 20–99 assemblies | 48.9% | 8.9% | 42.2% | 18 |
| 10–19 assemblies | 46.2% | 8.4% | 45.4% | 20 |
| 4–9 assemblies | 30.0% | 13.6% | 56.4% | 83 |
Percentage correct, incorrect, and indeterminate rates are averages across serovars stratified into groups by the number of genomes a serovar has in the reference database.