Table 1.
Ligand enrichments of three docking programs and 10 scoring functions.
| LBVS Approach | ROCE 0.5% a | ROCE 1% a | ROC AUC b | |
|---|---|---|---|---|
| docking programs | GOLD (Chemscore) | 0.00 | 0.00 | 0.63 |
| LibDock (LibScore) | 10.89 | 5.42 | 0.77 | |
| FRED (Chemgauss4) | 30.90 | 25.66 | 0.72 | |
| other scoring functions | Ludi_1 (DS c) | 0.00 | 0.00 | 0.39 |
| Ludi_2 (DS) | 0.00 | 0.00 | 0.40 | |
| Ludi_3 (DS) | 0.00 | 0.00 | 0.33 | |
| Ligscore1 (DS) | 0.00 | 2.57 | 0.56 | |
| Ligscore2 (DS) | 0.00 | 0.00 | 0.37 | |
| PLP1 (DS) | 5.15 | 7.70 | 0.50 | |
| PLP2 (DS) | 10.30 | 5.13 | 0.53 | |
| Jain (DS) | 5.15 | 2.57 | 0.40 | |
| PMF (DS) | 5.15 | 5.13 | 0.54 | |
| PMF04 (DS) | 0.00 | 5.13 | 0.52 | |
a The quotient of the true positive rate divided by the false positive rate at top-ranked 0.5% (for ROCE 0.5%) or 1% (for ROCE 1%) of binding decoys. A greater value indicates a better early recognition of a docking/scoring approach; b area under the ROC curve. A greater value represents a better overall enrichment; c Discovery Studio.