Table 2 Ablation studies of the AFR module and ESM-1v feature on the 5-fold cross-validation set of eSOL
From: HybridGCN for protein solubility prediction with adaptive weighting of multiple features
Methods | RMSE | \(R^2\) | Accuracy | Precision | Recall | F1 | AUC |
|---|---|---|---|---|---|---|---|
GCN | 0.231 | 0.483 | 0.779 | 0.775 | 0.693 | 0.732 | 0.866 |
ESM+GCN | 0.229 | 0.493 | 0.782 | 0.779 | 0.713 | 0.740 | 0.870 |
AFR+GCN | 0.230 | 0.488 | 0.781 | 0.778 | 0.715 | 0.742 | 0.871 |
ESM+AFR+GCN (HybridGCN) | 0.227 | 0.497 | 0.783 | 0.780 | 0.722 | 0.749 | 0.876 |