| 3D templates | SSM | SeqF_95 | SeqF_99 | SeqF_100 |
---|
Sensitivity |
0.897
| 0.724 |
0.931
| 0.862 | 0.552 |
Specificity |
1.000
|
0.933
| 0.600 | 0.667 |
0.933
|
PPV |
1.000
|
0.955
| 0.818 | 0.833 | 0.941 |
LSS-sensitivity | 0.200 | 0.267 |
0.533
|
0.467
| 0.133 |
- We calculated the sensitivity, specificity, and PPV for SeqFEATURE, 3D templates, and SSM using a random subset of 29 positive and 15 negative structures. Additionally, we calculated sensitivity for each method on a low structural similarity subset ('LSS-sensitivity') of the positive test set. Bold values indicate the top two performing methods for that row. SeqFEATURE performs relatively less well overall, but when structural similarity is reduced, SeqFEATURE again is the best performing method.