Skip to main content

Table 6 Performance comparison of our models on the independent BindingDB set

From: Multi-PLI: interpretable multi‐task deep learning model for unifying protein–ligand interaction datasets

Metric

Single-task on DUD-E

Single-task on Human

Single-task on C.elegans

Single-task on KIBA

Single-task on Davis

Single-task on PDBbind

Multi-task

Accuracy

0.521

0.534

0.532

0.480

0.532

0.587

0.580

AUC

0.501

0.512

0.547

0.432

–

–

0.613

  1. Best values are higlhlighted in bold
Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us