Table 7 Screening success of the general scoring functions trained with MLR, SMOreg and RF evaluated on the AKT2, KIT, and MK01 datasets from DUD-E. ac, dec and tot are the number of active, decoy compounds and the total number of molecules in the final dataset (i.e., compounds that were docked and rescored with DockThor and DockTScore, respectively). Only the top-scored protonation state of each compound according to each scoring function (SF) was kept.
From: New machine learning and physics-based scoring functions for drug discovery
Target | Metrics | General SFs | ||
|---|---|---|---|---|
MLR | SMOreg | RF | ||
AKT2 | AUC | 0.769 | 0.800 | 0.814 |
ac = 116 | EF1% (max = 60.414) | 24.166 | 15.535 | 13.809 |
dec = 6,892 | BEDROC20 | 0.421 | 0.378 | 0.379 |
tot = 7,008 | BEDROC100 | 0.394 | 0.288 | 0.269 |
KIT | AUC | 0.640 | 0.635 | 0.657 |
ac = 166 | EF1% (max = 63.934) | 3.016 | 2.413 | 5.428 |
dec = 10,447 | BEDROC20 | 0.148 | 0.146 | 0.176 |
tot = 10,613 | BEDROC100 | 0.063 | 0.043 | 0.090 |
MK01 | AUC | 0.786 | 0.766 | 0.745 |
ac = 78 | EF1% (max = 59.308) | 10.314 | 12.893 | 7.736 |
dec = 4,548 | BEDROC20 | 0.352 | 0.364 | 0.340 |
tot = 4.626 | BEDROC100 | 0.153 | 0.220 | 0.193 |
