Skip to main content

Table 5 Performance of CPI prediction with and without edge weights on human dataset

From: Sequence-based drug-target affinity prediction using weighted graph neural networks

Method

AUC

Precision

Recall

WGNN-DTA (GCN)

0.986

0.972

0.960

WGNN-DTA (GAT)

0.995

0.980

0.970

Without Weight (GCN)

0.967

0.927

0.869

Without Weight (GAT)

0.970

0.930

0.877

  1. The values in boldface represent the best prediction performances with the corresponding measures