Table 7 The impact of AlphaFold2 structural models on the performance of the GraphIdn
From: Identification of plant vacuole proteins by using graph neural network and contact maps
Model | Adjacency matrix | Acc (%) | Sn (%) | Sp (%) | MCC | ROC-AUC |
|---|---|---|---|---|---|---|
GraphIdn (With structural features) | With topology | 88.51 | 94.59 | 82.43 | 0.776 | 0.933 |
GraphIdn (Without structural features) | Random construction | 85.10 | 91.89 | 77.03 | 0.722 | 0.917 |
| Â | All 1 | 83.11 | 90.54 | 75.67 | 0.670 | 0.914 |
| Â | All 0 | 50.00 | 0.00 | 100.0 | 0.00 | 0.490 |