Open access
Sep 2026
RA-PLA: Retrieval-augmented graph convolutional networks for protein-ligand binding affinity prediction
A novel end-to-end framework that retrieves hard protein-ligand pairs—samples whose nearest neighbors have conflicting labels under manifold smoothness constraints—and integrates them via a semi-supervised graph convolutional network (GCN) that consistently outperforms state-of-the-art approaches.
Karim Abbasi, Hossein Banadkuki, R. Sepahvand et al.
· PLoS ONE · 0 citations