Table 2.
Overview of computational method parameters to assess the functionality of pharmacogenetic variants
| Conventional | ADME optimized | ||||||
|---|---|---|---|---|---|---|---|
| Algorithm | Category | Threshold | Sensitivity (%) | Specificity (%) | Threshold | Sensitivity (%) | Specificity (%) |
| SIFT | Functionality prediction algorithms | <0.05 | 80.7 | 54.2 | <0.0376 | 75.6 | 57.6 |
| PolyPhen-2 | >0.447 | 80.8 | 63 | >0.3841 | 83 | 61.6 | |
| LRT | <0.001 | 66.3 | 72.3 | <0.0025 | 77.3 | 65.2 | |
| MutationAssessor | >1.9 | 79 | 63.7 | >2.0566 | 74 | 67.8 | |
| FATHMM | <−1.5 | 18.2 | 81.9 | <0.486 | 69.9 | 27.1 | |
| FATHMM-MKL | >0.73 | 64.2 | 68 | >0.3982 | 77.4 | 63.3 | |
| PROVEAN | <−2.5 | 80.7 | 56.9 | <−3.286 | 72.2 | 72.2 | |
| VEST3 | >0.9 | 14.3 | 95.9 | >0.4534 | 67.6 | 78.8 | |
| GERP++ | Evolutionary conservation scores | >4.4 | 28.4 | 84.4 | >1.2482 | 84.2 | 47.6 |
| SiPhy | >12.17 | 32.1 | 78.2 | >7.2442 | 51.9 | 72.7 | |
| PhyloP (vertebrate) | NA | NA | NA | >0.5216 | 70.5 | 53.7 | |
| PhyloP (mammalian) | NA | NA | NA | >0.0461 | 77.4 | 49 | |
| PhastCons (vertebrate) | NA | NA | NA | >0.07 | 81.1 | 34.7 | |
| PhastCons (mammalian) | NA | NA | NA | >0.1872 | 67.4 | 49.7 | |
| CADD | Ensemble scores | >15 | 75.8 | 74.8 | >19.19 | 74.2 | 78.9 |
| DANN | >0.99 | 68.9 | 70.1 | >0.9688 | 85.8 | 54.4 | |
| MetaSVM | >0 | 43.4 | 86.3 | >−0.3371 | 51.6 | 78.1 | |
| MetaLR | >0.5 | 41.2 | 84.2 | >0.4039 | 52.2 | 76.7 | |
Sensitivity and specificity of each prediction method is shown for conventional disease dataset-based parameterization and ADME optimized parameters. Threshold values are in arbitrary units, values for sensitivity and specificity are provided in percentage (%)