Table 7.
DEBT's prediction of β-turn types on three different datasets.
| Dataset | β-turn type | MCC | Sensitivity (%) | Specificity (%) | Qtotal (%) | S (%) | AUC |
|---|---|---|---|---|---|---|---|
| GR426 | I | 0.36 (0.31) | 75.2 (67.5) | 78.9 (78.4) | 78.6 (77.9) | 30.1 (26.2) | 0.85 (0.82) |
| II | 0.29 (0.27) | 63.4 (65.0) | 88.3 (86.4) | 87.4 (85.7) | 23.1 (20.6) | 0.87 (0.86) | |
| IV | 0.27 (0.23) | 71.2 (63.4) | 71.5 (73.5) | 71.5 (72.5) | 20.4 (18.5) | 0.78 (0.76) | |
| VIII | 0.14 (0.10) | 68.7 (29.1) | 71.1 (89.8) | 71.1 (88.1) | 8.0 (7.7) | 0.77 (0.73) | |
| NS | 0.31 (0.28) | 18.0 (19.8) | 99.7 (99.4) | 97.6 (97.4) | 26.5 (26.1) | 0.81 (0.81) | |
| FA547 | I | 0.38 (0.31) | 71.6 (66.6) | 82.6 (79.5) | 81.6 (78.3) | 33.0 (26.0) | 0.85 (0.82) |
| II | 0.33 (0.27) | 63.0 (64.9) | 90.8 (86.8) | 89.8 (85.9) | 27.8 (20.9) | 0.88 (0.86) | |
| IV | 0.27 (0.24) | 69.8 (61.3) | 73.3 (75.6) | 73.0 (74.3) | 21.0 (19.2) | 0.79 (0.77) | |
| VIII | 0.14 (0.10) | 47.8 (28.4) | 84.4 (90.2) | 83.4 (88.5) | 9.5 (7.9) | 0.77 (0.73) | |
| NS | 0.37 (0.28) | 21.1 (21.2) | 99.7 (99.2) | 97.7 (97.2) | 31.2 (26.3) | 0.84 (0.82) | |
| FA823 | I | 0.39 (0.30) | 70.6 (64.3) | 84.2 (80.7) | 83.0 (79.3) | 34.5 (26.0) | 0.86 (0.82) |
| II | 0.33 (0.28) | 62.7 (65.1) | 91.2 (87.2) | 90.2 (86.4) | 27.9 (21.1) | 0.88 (0.86) | |
| IV | 0.27 (0.23) | 68.3 (58.6) | 74.4 (77.1) | 73.9 (75.5) | 21.0 (18.9) | 0.79 (0.76) | |
| VIII | 0.14 (0.08) | 42.2 (12.4) | 87.2 (96.6) | 86.1 (94.5) | 9.4 (7.3) | 0.77 (0.72) | |
| NS | 0.38 (0.29) | 23.6 (24.2) | 99.7 (98.9) | 97.7 (97.0) | 33.9 (27.9) | 0.85 (0.83) |
In the parentheses is the prediction using PSSM-only input without predicted dihedral angles or secondary structure. Notably, there is improvement in the predictive performance when the input vector is augmented by predicted dihedral angles and secondary structures.