Table 2.
Accuracy comparison
| Ibis Test |
B.pertussis |
PhiX174 (AYB) |
|||||
|---|---|---|---|---|---|---|---|
| Perfect reads | Error rate (%) | Perfect reads | Error rate (%) | Perfect reads | Error rate (%) | ||
| Bustard | 99 834 | 1.45 | 1 557 963 | 2.01 | 24 478 | 0.49 | |
| AYB | 133 537 | 0.73 | 2 304 005 | 1.26 | 26 878 | 0.38 | |
| BlindCall slow | 110 951 | 1.12 | 1 902 621 | 1.61 | 25 144 | 0.45 | |
| BlindCall fast | 105 312 | 1.26 | 1 856 286 | 1.66 | 24 740 | 0.47 | |
| Time | Slow | 0.08/0.3/1 | 0.11/6/10 | 0.15/14/22 | |||
| Fast | 0.08/0.1/1 | 0.11/3/8 | 0.15/7/16 | ||||
Accuracy for Bustard, AYB and BlindCall on various datasets. BlindCall was able to produce comparable accuracy to state-of-the-art base callers at significantly faster computational time. All methods improve on Bustard base calls. Run times for BlindCall are reported as (training time/processing time/total time in minutes) where the total time includes reading intensity data from disk and writing base-calls to disk.