Table 2.
Estimated threshold ranges for all tools in this study corresponding to the four pathogenic and four benign intervals
| Method |
Benign (BP4) |
Pathogenic (PP3) |
||||||
|---|---|---|---|---|---|---|---|---|
| Very Strong | Strong | Moderate | Supporting | Supporting | Moderate | Strong | Very Strong | |
| BayesDel | – | – | ≤ −0.36 | (−0.36, −0.18] | [0.13, 0.27) | [0.27, 0.50) | ≥0.50 | – |
| CADD | – | ≤0.15 | (0.15, 17.3] | (17.3, 22.7] | [25.3, 28.1) | ≥28.1 | – | – |
| EA | – | – | ≤0.069 | (0.069, 0.262] | [0.685, 0.821) | ≥0.821 | – | – |
| FATHMM | – | – | ≥4.69 | [3.32, 4.69) | (−5.04, −4.14] | ≤ −5.04 | – | – |
| GERP++ | – | – | ≤ −4.54 | (−4.54, 2.70] | – | – | – | – |
| MPC | – | – | – | – | [1.360, 1.828) | ≥1.828 | – | – |
| MutPred2 | – | ≤0.010 | (0.010, 0.197] | (0.197, 0.391] | [0.737, 0.829) | [0.829, 0.932) | ≥0.932 | – |
| PhyloP | – | – | ≤0.021 | (0.021, 1.879] | [7.367, 9.741) | ≥9.741 | – | – |
| PolyPhen2 | – | – | ≤0.009 | (0.009, 0.113] | [0.978, 0.999) | ≥0.999 | – | – |
| PrimateAI | – | – | ≤0.362 | (0.362, 0.483] | [0.790, 0.867) | ≥0.867 | – | – |
| REVEL | ≤0.003 | (0.003, 0.016] | (0.016, 0.183] | (0.183, 0.290] | [0.644, 0.773) | [0.773, 0.932) | ≥0.932 | – |
| SIFT | – | – | ≥0.327 | [0.080, 0.327) | (0, 0.001] | 0 | – | – |
| VEST4 | – | – | ≤0.302 | (0.302, 0.449] | [0.764, 0.861) | [0.861, 0.965) | ≥0.965 | – |
A “–” implies that the given tool did not meet the posterior probability (likelihood ratio) threshold. See Table S1 for comprehensive results that include point estimates and one-sided confidence intervals. Intervals follow standard mathematical notation in which “(” and “)” indicate exclusion of the end value and “[” and “]” indicate inclusion of the end value