Table 1 Comparison of the performance of PEDLA with that of existing methods.
From: PEDLA: predicting enhancers with a deep learning-based algorithmic framework
PEDLA | RFECS | CSI-ANN | DELTA | ChromHMM | Segway | PEDLA (all features) | |||
|---|---|---|---|---|---|---|---|---|---|
Number of prediction | 22691 | 75084 | 30173 | 112044 | 26869 | 131698 | 20689 | ||
Performance metrics | Accuracy | 96.30% | 93.67% | 95.58% | 87.78% | 94.03% | 91.01% | 97.65% | |
Sensitivity | 95.72% | 64.19% | 65.50% | 73.56% | 37.67% | 12.89% | 96.16% | ||
Specificity | 96.37% | 97.89% | 98.63% | 89.84% | 99.75% | 98.94% | 97.80% | ||
GM | 96.02% | 79.26% | 80.34% | 81.29% | 61.30% | 35.71% | 96.97% | ||
F1-score | 83.01% | 71.71% | 73.06% | 60.40% | 53.74% | 20.90% | 88.31% | ||
Validation rate | DHS | 40.68% | 31.85% | 30.65% | 12.25% | 38.86% | 40.61% | 42.29% | |
P300 | 15.25% | 7.26% | 10.83% | 1.57% | 9.89% | 3.52% | 16.82% | ||
TFs | 28.89% | 17.71% | 19.72% | 5.75% | 19.14% | 6.42% | 32.37% | ||
Misclassification rate | 7.53% | 3.09% | 16.46% | 3.01% | 6.42% | 14.53% | 6.59% | ||
