Table 1:
Database | Approach | Threshold | Threshold category | Recall | Precision | F-score | FDR (%) |
---|---|---|---|---|---|---|---|
eggNOG | Assembly-free | BTS 120 | Strict (TH) | 0.9880 | 0.9540 | 0.9707 | 4.5977 |
CoMW | 1.00E−15 | Strict (TH) | 0.9851 | 0.9939 | 0.9895 | 0.6006 | |
CAZy | Assembly-free | BTS 110 | Strict (TH) | 0.3510 | 0.5325 | 0.4231 | 46.7433 |
CoMW | 1.00E−08 | Medium (TM) | 0.8131 | 0.7759 | 0.7940 | 22.4096 | |
NCycDB | Assembly-free | BTS 150 | Strict (TH) | 0.1666 | 0.0581 | 0.0862 | 94.1860 |
CoMW | 1.00E−14 | Strict (TH) | 0.6666 | 0.8333 | 0.7407 | 16.6666 |
Full table for both approaches and databases can be seen in Tables S1–S3. Boldface emphasizes better precision, recall, F-score, and FDR in each database between both approaches.