Table 2 Performance of proposed and baseline models in task of drug-blind prediction.
From: Drug discovery and mechanism prediction with explainable graph neural networks
Method | Conv type | RMSE (\(\downarrow\)) | PCC (\(\uparrow\)) | \(\mathbf {R^2}\) (\(\uparrow\)) |
|---|---|---|---|---|
tCNN11 | CNN | 0.056 ± 0.001 | 0.356 ± 0.019 | 0.027 ± 0.010 |
GraphDRP25 | GCN | 0.063 ± 0.002 | 0.450 ± 0.026 | 0.153 ± 0.048 |
GAT | 0.071 ± 0.003 | 0.351 ± 0.165 | -0.041 ± 0.045 | |
TGSA (exp)27 | GraphSAGE | 2.809 ± 0.035 | 0.329 ± 0.058 | 0.026 ± 0.078 |
XGDP | GCN | 0.056 ± 0.000 | 0.400 ± 0.016 | 0.048 ± 0.015 |
GAT | 0.053 ± 0.001 | 0.448 ± 0.036 | 0.149 ± 0.052 | |
GAT_E | 0.052 ± 0.003 | 0.505 ± 0.090 | 0.164 ± 0.043 | |
GATv2_E | 0.055 ± 0.002 | 0.442 ± 0.041 | 0.058 ± 0.024 | |
RGCN | 0.055 ± 0.001 | 0.405 ± 0.031 | 0.063 ± 0.045 | |
RGAT | 0.055 ± 0.002 | 0.257 ± 0.061 | 0.063 ± 0.060 |
