Table 3.
Performance comparison of feature combinations on the benchmark dataset (HB34) using GTB.
| Methods | Features | SPE | SEN | PRE | ACC | F1 | MCC | AUC |
|---|---|---|---|---|---|---|---|---|
| GTB | ASA + PSSM | 0.708 | 0.705 | 0.642 | 0.707 | 0.663 | 0.410 | 0.761 |
| PSSM + Solvent exposure | 0.671 | 0.718 | 0.617 | 0.691 | 0.656 | 0.385 | 0.760 | |
| Blosum62 + Solvent exposure | 0.664 | 0.699 | 0.606 | 0.679 | 0.640 | 0.359 | 0.734 | |
| ASA + Solvent exposure | 0.674 | 0.695 | 0.612 | 0.683 | 0.642 | 0.366 | 0.728 | |
| Phy+Solvent exposure | 0.664 | 0.696 | 0.605 | 0.677 | 0.639 | 0.357 | 0.728 | |
| ASA + Blosum62 | 0.658 | 0.651 | 0.585 | 0.656 | 0.608 | 0.307 | 0.718 | |
| ASA + Phy | 0.669 | 0.644 | 0.590 | 0.658 | 0.607 | 0.311 | 0.717 | |
| Phy + PSSM | 0.629 | 0.650 | 0.566 | 0.638 | 0.597 | 0.277 | 0.683 | |
| PSSM + Blosum62 | 0.619 | 0.655 | 0.560 | 0.635 | 0.595 | 0.271 | 0.679 | |
| Phy + Blosum62 | 0.593 | 0.590 | 0.520 | 0.592 | 0.541 | 0.183 | 0.639 | |
| Combined (all features) | 0.717 | 0.656 | 0.727 | 0.719 | 0.681 | 0.439 | 0.787 |