Table 1.
Comparison of DeepEnhancerPPO with 24 state-of-the-art models on the independent datasets for five metrics on the first enhancer classification task: category classification.
| Model | SN | SP | ACC | MCC | AUC |
|---|---|---|---|---|---|
| iEnhancer-2L [15] | 0.7100 | 0.7500 | 0.7300 | 0.4604 | 0.8062 |
| EnhancerPred [16] | 0.7350 | 0.7450 | 0.7400 | 0.4800 | 0.8013 |
| iEnhancer-EL [34] | 0.7100 | 0.7850 | 0.7475 | 0.4964 | 0.8173 |
| iEnhancer-5Step [35] | 0.8200 | 0.7600 | 0.7900 | 0.5800 | - |
| DeployEnhancer [36] | 0.7550 | 0.7600 | 0.7550 | 0.5100 | 0.7704 |
| iEnhancer-ECNN [18] | 0.7520 | 0.7850 | 0.7690 | 0.5370 | 0.8320 |
| EnhancerP-2L [37] | 0.7810 | 0.8105 | 0.7950 | 0.5907 | - |
| iEnhancer-CNN [38] | 0.7825 | 0.7900 | 0.7750 | 0.5850 | - |
| iEnhancer-XG [39] | 0.7400 | 0.7750 | 0.7575 | 0.5150 | - |
| Enhancer-DRRNN [40] | 0.7330 | 0.8010 | 0.7670 | 0.5350 | 0.8370 |
| Enhancer-BERT [22] | 0.8000 | 0.7120 | 0.7560 | 0.5140 | - |
| iEnhancer-RF [41] | 0.7850 | 0.8100 | 0.7975 | 0.5952 | 0.8600 |
| spEnhancer [42] | 0.8300 | 0.7150 | 0.7725 | 0.5793 | 0.8235 |
| iEnhancer-EBLSTM [43] | 0.7550 | 0.7950 | 0.7720 | 0.5340 | 0.8350 |
| iEnhancer-GAN [44] | 0.8110 | 0.7580 | 0.7840 | 0.5670 | - |
| piEnhPred [45] | 0.8250 | 0.7840 | 0.8040 | 0.6099 | - |
| iEnhancer-RD [44] | 0.8100 | 0.7650 | 0.7880 | 0.5760 | 0.8440 |
| iEnhancer-MFGBDT [46] | 0.7679 | 0.7955 | 0.7750 | 0.5607 | - |
| Enhancer-LSTMAtt [19] | 0.7950 | 0.8150 | 0.8050 | 0.6101 | 0.8588 |
| iEnhancer-Deep [47] | 0.8150 | 0.6700 | 0.7402 | 0.4902 | - |
| Rank-GAN [48] | 0.7487 | 0.7563 | 0.7525 | 0.5051 | 0.8062 |
| iEnhancer-DCLA [49] | 0.7800 | 0.7850 | 0.7825 | 0.5650 | 0.8269 |
| iEnhancer-BERT [50] | - | - | 0.7930 | 0.5850 | 0.8440 |
| ADH-Enhancer [51] | 0.8420 | 0.8700 | 0.8430 | 0.6860 | - |
| DeepEnhancerPPO (Ours) | 0.9000 | 0.9200 | 0.9100 | 0.8202 | 0.9585 |