Table 2 Comparing Performance of DeepLPI, DeepCDA and DeepDTA on the internal independent testing set from the BindingDB data.

From: DeepLPI: a novel deep learning-based model for protein–ligand interaction prediction for drug repurposing

 

AUROC

Sensitivity

Specificity

PPV

NPV

DeepLPI

0.925

0.855

0.862

0.306

0.988

DeepCDA

0.912

0.766

0.896

0.342

0.982

DeepDTA

0.909

0.865

0.795

0.221

0.989

Unseen testsets combined

This work

0.791

0.661

0.789

0.132

0.980

DeepCDA

0.741

0.511

0.813

0.495

0.823

  1. Significant values are in bold.