| Model condition | PR-AUC (%) |
|---|
| ClustalW | FSA |
|---|
| GUIDANCE2 | SERES + GUIDANCE2 | Pairwise t-test corrected q-value | GUIDANCE2 | SERES + GUIDANCE2 | Pairwise t-test corrected q-value |
|---|
| 10.A | 95.37 | 95.78 | \(2.8 \times 10^{-3}\) | 96.36 | 96.55 | \(8.6 \times 10^{-3}\) |
| 10.B | 92.30 | 92.95 | \(8.2 \times 10^{-4}\) | 95.40 | 95.87 | \(4.9 \times 10^{-3}\) |
| 10.C | 89.36 | 91.23 | \(1.7 \times 10^{-4}\) | 95.32 | 96.06 | \(2.7 \times 10^{-3}\) |
| 10.D | 88.53 | 90.45 | \(8.8 \times 10^{-5}\) | 96.21 | 96.87 | \(2.1 \times 10^{-3}\) |
| 10.E | 73.96 | 76.50 | \(8.2 \times 10^{-4}\) | 90.23 | 92.51 | \(8.6 \times 10^{-3}\) |
| Model condition | ROC-AUC (%) |
|---|
| ClustalW | FSA |
|---|
| GUIDANCE2 | SERES + GUIDANCE2 | DeLong et al. test corrected q-value | GUIDANCE2 | SERES + GUIDANCE2 | DeLong et al. test corrected q-value |
|---|
| 10.A | 96.99 | 97.23 | \(<10^{-10}\) | 80.85 | 81.61 | \(<10^{-10}\) |
| 10.B | 96.64 | 96.94 | \(<10^{-10}\) | 81.31 | 82.89 | \(<10^{-10}\) |
| 10.C | 96.27 | 96.88 | \(<10^{-10}\) | 84.48 | 86.56 | \(<10^{-10}\) |
| 10.D | 95.78 | 96.65 | \(<10^{-10}\) | 88.63 | 90.37 | \(<10^{-10}\) |
| 10.E | 89.84 | 90.80 | \(<10^{-10}\) | 89.10 | 90.83 | \(<10^{-10}\) |
- Input MSAs in these experiments were estimated using either ClustalW [13] or FSA [2] (MAFFT was used to estimate input MSAs throughout the rest of our study.) Results are shown for model conditions 10.A through 10.E (named in order of generally increasing sequence divergence). The best AUC for each pairwise method comparison on a model condition is shown in italics. Otherwise, table layout and description are identical to Table 6
