Table 2.
Median AUROC and AUPR for R8 when all networks are used (All) compared to when default (-default), co-expression (-coexp), co-localization (-coloc), genetic interaction (-gi), pathway (-path), physical (protein) interaction (-pi), predicted (-predict), and shared protein domain (-spd) networks are removed, respectively.
| All | -default | -coexp | -coloc | -gi | -path | -pi | -predict | -spd | |
|---|---|---|---|---|---|---|---|---|---|
| R8 | |||||||||
| Total edges | 1.64 × 108 | 1.50 × 108 | 6.94 × 106 | 1.63 × 108 | 1.59 × 108 | 1.64 × 108 | 1.63 × 108 | 1.63 × 108 | 1.63 × 108 |
| Edges removed from all | 0 | 1.37 × 107 | 1.57 × 108 | 4.87 × 105 | 4.85 × 106 | 1.16 × 105 | 2.75 × 105 | 1.99 × 105 | 1.02 × 106 |
| R8: GO TERM SIZE = 3–10 | |||||||||
| Median AUROC | 0.694 | 0.685 | 0.675 | 0.694 | 0.695 | 0.692 | 0.694 | 0.694 | 0.688 |
| 95% CI | ±0.311 | ±0.309 | ±0.332 | ±0.311 | ±0.310 | ±0.311 | ±0.309 | ±0.310 | ±0.313 |
| % difference from all | −1.3% | −2.8% | −0.1% | 0.1% | −0.4% | −0.1% | −0.1% | −0.9% | |
| versus all (p-value) | *2.17 × 10−3 | *1.43 × 10−23 | *5.04 × 10−67 | *4.23 × 10−109 | *1.46 × 10−4 | *5.86 × 10−13 | *4.05 × 10−17 | 7.05 × 10−1 | |
| R8: GO TERM SIZE = 11–300 | |||||||||
| Median AUROC | 0.890 | 0.866 | 0.864 | 0.889 | 0.890 | 0.887 | 0.887 | 0.890 | 0.880 |
| 95% CI | ±0.195 | ±0.206 | ±0.224 | ±0.196 | ±0.195 | ±0.197 | ±0.199 | ±0.196 | ±0.199 |
| % difference from all | −2.8% | −2.9% | −0.1% | 0.0% | −0.3% | −0.4% | 0.0% | −1.1% | |
| versus all (p-value) | *0 | *6.91 × 10−183 | 1.91 × 10−1 | *8.24 × 10−20 | *1.54 × 10−22 | *1.01 × 10−27 | 8.83 × 10−1 | *1.88 × 10−219 | |
| R8: GO TERM SIZE = 3–10 | |||||||||
| Median AUPR | 0.019 | 0.011 | 0.025 | 0.019 | 0.020 | 0.019 | 0.016 | 0.018 | 0.017 |
| 95% CI | ±0.409 | ±0.377 | ±0.406 | ±0.409 | ±0.408 | ±0.407 | ±0.392 | ±0.411 | ±0.404 |
| % difference from all | −44.8% | 29.8% | 0.0% | 3.4% | −4.0% | −16.4% | −4.8% | −14.0% | |
| versus all (p-value) | *2.43 × 10−9 | *5.94 × 10−15 | *6.98 × 10−61 | *5.10 × 10−83 | *8.71 × 10−13 | *1.97 × 10−11 | *3.40 × 10−24 | *1.82 × 10−4 | |
| R8: GO TERM SIZE = 11–300 | |||||||||
| Median AUPR | 0.220 | 0.186 | 0.209 | 0.219 | 0.220 | 0.218 | 0.206 | 0.218 | 0.215 |
| 95% CI | ±0.528 | ±0.527 | ±0.530 | ±0.528 | ±0.528 | ±0.525 | ±0.528 | ±0.528 | ±0.527 |
| % difference from all | −15.6% | −5.0% | −0.4% | 0.1% | −0.6% | −6.2% | −0.9% | −2.2% | |
| versus all (p-value) | *1.14 × 10−181 | *4.89 × 10−9 | 3.20 × 10−1 | *3.63 × 10−38 | *3.92 × 10−16 | *1.20 × 10−43 | *3.40 × 10−24 | *1.82 × 10−4 | |
The number of edges removed for each analysis is also listed, as well as total edges used during assessment. The Wilcoxon rank sum test was used to compute the listed p-values, where significant values (p < 0.01) are marked with an asterisk.