Table 2. Performance evaluation and comparison using independent test set (NovelVar) of iFish, SIFT, MutationAssessor, PolyPhen2 trained by HumDiv and HumVar, CADD, MutationTaster2, FATHMM and Condel.
| Accuracy(%) | FDR(%) | Sensitivity(%) | Specificity(%) | MCC | |
|---|---|---|---|---|---|
| iFish | 76.99 | 28.16 | 81.06 | 73.60 | 0.54 |
| SIFT | 68.15 | 38.81 | 86.00 | 52.66 | 0.41 |
| PolyPhen2 HDIV | 69.81 | 37.18 | 88.22 | 53.37 | 0.44 |
| PolyPhen2 HVAR | 72.84 | 32.71 | 82.56 | 64.15 | 0.47 |
| Mutation Assessor | 72.70 | 31.23 | 75.03 | 70.69 | 0.46 |
| Mutation Taster2 | 71.78 | 37.81 | 91.08 | 56.68 | 0.49 |
| CADD | 76.20 | 28.76 | 79.75 | 73.25 | 0.53 |
| FATHMM | 73.07 | 25.77 | 62.66 | 81.78 | 0.45 |
| Condel | 76.74 | 27.41 | 80.77 | 73.20 | 0.54 |
FDR: False Discovery Rate. MCC: Matthews Correlation Coefficient.