Table 1.
GESPA pathogenicity classifier accuracy compared to other software using humsavar test set
| Software | Sensitivity (%)a | Specificity (%)b | Balanced accuracy (%)c | ROC Curve AUC |
|---|---|---|---|---|
| GESPA (humsavar cross validation) | 96.41 | 79.49 | 87.95 | 0.936 |
| GESPA (humvar training set) | 96.31 | 79.23 | 87.78 | 0.932 |
| Polyphen 2 | 88.68 | 62.45 | 75.56 | 0.847 |
| SIFT | 85.03 | 68.95 | 76.99 | 0.854 |
| PROVEAN | 78.39 | 79.11 | 78.75 | 0.848 |
| Mutation Assessor | 85.29 | 71.02 | 78.15 | 0.848 |
aSensitivity = TP/(TP + FN)
bSpecificity = TN/(TN + FP)
cBalanced Accuracy = (Sensitivity + Specificity)/2