Skip to main content

Table 2 Performance of algorithms to classify drug–target interactions

From: GraphDTI: A robust deep learning predictor of drug-target interactions from multiple heterogeneous data

Algorithm

GraphDTI dataset

PubChem Bioassay dataset

Random-split

Cluster-based

GraphDTI

0.999 ±0.0004

0.996 ±0.0036

0.939

EnsemDT

0.924 ±0.0903

0.824 ±0.0972

0.597

EnsemKRR

0.977 ±0.0029

0.885 ±0.0365

0.488

RLS-Kron

0.976 ±0.0035

0.834 ±0.0393

0.465

Vina

0.534 ±0.0044

0.551 ±0.0372

–

  1. The Area Under the Curve (AUC) measures the classification performance against the GraphDTI dataset, cross-validated with random-split and cluster-based protocols, and the PubChem Bioassay dataset containing unseen data
Back to article page

Journal of Cheminformatics

ISSN: 1758-2946

Contact us