Table 2 The performance comparison between MMCLKin and MMAtt-DTA on seven target superfamilies
From: Enhancing kinase-inhibitor activity and selectivity prediction through contrastive learning
Datasets | Model | Spearman↑ | RMSE↓ | CI↑ |
|---|---|---|---|---|
Enzyme | MMAtt-DTA | 0.720 | 0.509 | 0.866 |
MMCLKin | 0.727 | 0.479 | 0.870 | |
Epigenetic regulator | MMAtt-DTA | 0.470 | 0.560 | 0.811 |
MMCLKin | 0.500 | 0.591 | 0.816 | |
GPCR | MMAtt-DTA | 0.878 | 0.679 | 0.865 |
MMCLKin | 0.881 | 0.667 | 0.867 | |
Ion channel | MMAtt-DTA | 0.877 | 0.644 | 0.875 |
MMCLKin | 0.878 | 0.596 | 0.875 | |
Kinase | MMAtt-DTA | 0.873 | 0.625 | 0.861 |
MMCLKin | 0.887 | 0.577 | 0.869 | |
Nuclear receptor | MMAtt-DTA | 0.722 | 0.779 | 0.822 |
MMCLKin | 0.711 | 0.799 | 0.820 | |
Transporter | MMAtt-DTA | 0.856 | 0.696 | 0.848 |
MMCLKin | 0.874 | 0.616 | 0.860 |
