Skip to main content

Table 6 Comparing the effect of clustering on the prediction performance of DEEPLYESSENTIAL on the GP+GN dataset

From: DeeplyEssential: a deep neural network for predicting essential genes in microbes

Metric

Non-clustered

Clustered

Difference (%)

AUC

0.842

0.786

7.12%

Sensitivity

0.801

0.780

2.69%

Specificity

0.721

0.646

11.60%

PPV

0.749

0.688

8.86%

Accuracy

0.762

0.713

6.87%