Table 5. Combined performance statistics of SVM classifiers employing solo features and hybrid approaches in predicting O-glycosites (using balanced dataset).
Feature | Sensitivity (%) | Specificity (%) | Accuracy (%) | MCC (%) | AUC (%) |
CPP | 68.10 | 72.41 | 70.26 | 0.41 | 0.74071 |
CPP+SS | 70.69 | 71.55 | 71.12 | 0.42 | 0.75743 |
CPP+ASA | 67.24 | 75.00 | 71.12 | 0.42 | 0.75780 |
CPP+SS+ASA | 72.41 | 75.00 | 73.71 | 0.47 | 0.76955 |
BPP | 66.38 | 67.24 | 66.81 | 0.34 | 0.73023 |
BPP+SS | 69.83 | 68.10 | 68.97 | 0.38 | 0.74160 |
BPP+ASA | 77.59 | 61.21 | 69.40 | 0.39 | 0.71143 |
BPP+SS+ASA | 65.52 | 72.41 | 68.97 | 0.38 | 0.73766 |
PPP | 75.00 | 71.55 | 73.28 | 0.47 | 0.81250 |
PPP+SS | 73.28 | 73.28 | 73.28 | 0.47 | 0.76806 |
PPP+ASA | 74.14 | 71.55 | 72.84 | 0.46 | 0.77341 |
PPP+SS+ASA | 77.59 | 69.83 | 73.71 | 0.48 | 0.76925 |
Footnotes: BPP- Binary profile of patterns, CPP- Composition profile of patterns, PPP- PSSM profile of patterns, MCC- Matthews correlation coefficient, AUC- Area under curve, SS-secondary structure and ASA- Accessible surface area.