TABLE 1.
Accuracy of serotype predictions
| Result | No. of genomes (% of total) |
||
|---|---|---|---|
| Reads mapping, CDC strains | Reads mapping, GenomeTrakr strains | Assembled genomes | |
| Expected serotypea | 304 (98.7) | 3,061 (92.6) | 324 (91.5) |
| Unexpected serotype | 2 (0.65) | 205 (6.2)b | 11 (3.1)b |
| Partial or no serotypec | 2 (0.65) | 40 (1.2) | 19 (5.4) |
| Total tested | 308 | 3306 | 354 |
a
The identification of the predicted serotype was considered correct when the serotype antigens detected corresponded to the antigens detected by conventional methods. See Text S1 in the supplemental material for a discussion of interpretation of serotype results. For GenomeTrakr and genome assembly datasets, serotype prediction in consensus with annotated serotype was considered correct.
b
Numbers represent serotype predictions inconsistent with the annotated serotype; the accuracy of the annotated serotype is unknown.
c
Some or all of the expected serotype determinants were not detected.