Table 2. Comparison with other predictors on the CamSol dataset.
Trevino | Miklos | Tan | Dudgeon | Total | Accuracy | |
---|---|---|---|---|---|---|
SolPro | 15 / 22 | 3 / 3 | 1 / 1 | 21 / 30 | 40 / 56 | 71.4 |
PROSO II | 16 / 22 | 3 / 3 | 1 / 1 | 12 / 30 | 32 / 56 | 57.1 |
CamSol | 22 / 22 | 3 / 3 | 1 / 1 | 28 / 30 | 54 / 56 | 96.4 |
SODA | 22 / 22 | 3 / 3 | 1 / 1 | 30 / 30 | 56 / 56 | 100.0 |
SODA is compared to three published methods. (20). The dataset is the same used in the recent CamSol paper (21) and includes 19 proteins and 56 variants from four publications: Trevino (15), Miklos (30), Tan (31) and Dudgeon (32). Accuracy is calculated as the percentage of correct predictions over the dataset size.