Table 3 Performance of models 2, 3, 4 trained in schemes 3, 4, 5, 6 and tested in schemes 1, 2, 5, 6 on the PDBbind v2007 benchmark
From: Correcting the impact of docking pose generation error on binding affinity prediction
Model | Training | Test | RMSE | SD | Rp | Rs |
|---|---|---|---|---|---|---|
2 (MLR::Vina) | 3 | 1 | 1.89 | 1.85 | 0.629 | 0.675 |
3 (RF::Vina) | 3 | 1 | 1.76 | 1.73 | 0.691 | 0.694 |
4 (RF::VinaElem) | 3 | 1 | 1.58 | 1.45 | 0.795 | 0.792 |
2 (MLR::Vina) | 3 | 2 | 1.88 | 1.85 | 0.630 | 0.661 |
3 (RF::Vina) | 3 | 2 | 1.72 | 1.68 | 0.711 | 0.714 |
4 (RF::VinaElem) | 3 | 2 | 1.57 | 1.45 | 0.793 | 0.780 |
2 (MLR::Vina) | 4 | 1 | 1.93 | 1.93 | 0.589 | 0.648 |
3 (RF::Vina) | 4 | 1 | 1.81 | 1.80 | 0.656 | 0.669 |
4 (RF::VinaElem) | 4 | 1 | 1.63 | 1.53 | 0.769 | 0.769 |
2 (MLR::Vina) | 4 | 2 | 1.94 | 1.93 | 0.589 | 0.636 |
3 (RF::Vina) | 4 | 2 | 1.79 | 1.75 | 0.682 | 0.686 |
4 (RF::VinaElem) | 4 | 2 | 1.63 | 1.53 | 0.769 | 0.762 |
2 (MLR::Vina) | 5 | 5 | 1.90 | 1.89 | 0.609 | 0.641 |
3 (RF::Vina) | 5 | 5 | 1.74 | 1.70 | 0.700 | 0.699 |
4 (RF::VinaElem) | 5 | 5 | 1.65 | 1.55 | 0.760 | 0.754 |
2 (MLR::Vina) | 6 | 6 | 1.86 | 1.83 | 0.640 | 0.670 |
3 (RF::Vina) | 6 | 6 | 1.73 | 1.69 | 0.707 | 0.707 |
4 (RF::VinaElem) | 6 | 6 | 1.60 | 1.49 | 0.780 | 0.769 |