Table 2 Representative deep-learning PLBAP models and their scoring performances
From: Structure-based, deep-learning models for protein-ligand binding affinity prediction
Model (ID) | Training (PDBbind Refined Set) | Parameter-tuning (PDBbind Core Set) | Test1 (CSAR-HiQ Set 1) | Test2 (CSAR-HiQ Set 2) | ||||
|---|---|---|---|---|---|---|---|---|
PC | RMSE | PC | RMSE | PC | RMSE | PC | RMSE | |
\(\mathbf {T_{ACNN}}\) (\(M_5\)) | 0.5189 | 1.7564 | 0.5692 | 1.7939 | 0.5596 | 1.9749 | 0.6804 | 1.6298 |
\(\mathbf {T_{IMC-CNN}}\) (\(M_9\)) | 0.7851 | 1.2607 | 0.7843 | 1.4807 | 0.6365 | 1.8011 | 0.6123 | 1.7329 |
\(\mathbf {T_{Grid-CNN}}\) (\(M_{12}\)) | 0.9224 | 0.8939 | 0.9235 | 1.0079 | 0.5531 | 1.9451 | 0.684 | 1.6373 |
\(\mathbf {T_{Graph-GCN}}\) (\(M_{26}\)) | 0.6403 | 1.5178 | 0.6969 | 1.6733 | 0.6706 | 1.7414 | 0.737 | 1.5098 |