Table 5.
Realistic synthetic data: classification-based results
|
Conservation |
Known width (w∗) |
MCOIN (w∗ ± 4) |
E-values (w∗ ± 4) |
||||||
|---|---|---|---|---|---|---|---|---|---|
| (mean bits/col) | sSn | sPPV | AUC | sSn | sPPV | AUC | sSn | sPPV | AUC |
|
2.00 |
0.84 |
0.25 |
0.99 |
0.93 |
0.42 |
1.00 |
0.91 |
0.79 |
0.99 |
|
1.49 |
0.26 |
0.07 |
0.98 |
0.28 |
0.15 |
0.99 |
0.21 |
0.45 |
0.98 |
|
1.08 |
0.02 |
0.01 |
0.96 |
0.01 |
0.01 |
0.96 |
0.01 |
0.23 |
0.96 |
|
0.76 |
0.00 |
0.00 |
0.94 |
0.00 |
0.00 |
0.93 |
0.00 |
0.12 |
0.94 |
| 0.51 | 0.00 | 0.00 | 0.93 | 0.00 | 0.00 | 0.93 | 0.00 | 0.09 | 0.93 |
Mean site-level sensitivity (sSn), positive predictive value (sPPV) and area under the ROC curve (AUC) for five collections of realistic synthetic data at varying levels of motif conservation. In these tests, the motif discovery algorithm was allowed to run as it would normally.