RA-PLA: Retrieval-augmented graph convolutional networks for protein-ligand binding affinity prediction
- 1Mosaheb Institute for Mathematical Research, Kharazmi University, Tehran, Iran
- 2Computer Engineering and Information Technology Department, University of Qom, Qom, Iran
- 3Department of Artificial Intelligence and Data Science, College of Science and Technology, University of Human Development, Sulaymaniyah, Kurdistan Region ofIraq
Abstract
Protein-ligand binding affinity (PLA) prediction is essential for computational drug discovery, yet existing deep learning methods typically employ a single globally trained model that fails to adapt to individual query samples. In this paper, we propose 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). For each query, we automatically construct a graph where nodes represent protein-ligand pairs and edges encode pairwise similarities. Our model jointly learns joint descriptors, graph topology, and the GCN predictor. During inference, we fine-tune the model per query using its retrieved hard neighbors.We evaluate on four benchmarks: PDBbind, Davis, KIBA, and BindingDB. Our method consistently outperforms state-of-the-art approaches. On Davis, we achieve a Concordance Index (CI) of 0.952 ± 0.005 (1.6% improvement over NerLTR-DTA) and AUPR of 0.815 ± 0.005 (7.6% improvement). On KIBA, we achieve CI of 0.916 ± 0.002 and AUPR of 0.871 ± 0.005, outperforming DeepCDA by 2.7% and 5.9%, respectively. On PDBbind, our CI of 0.894 ± 0.002 represents an 11.4% improvement over PLA-MoRe. On BindingDB, we achieve AUPR of 0.518 ± 0.009 (5.9% improvement over DeepCDA). In cold-target generalization across four unseen protein families, we achieve average AUPR improvements of 15.5%. Paired t-tests confirm statistical significance (p < 0.05), and ablation studies validate that both hard sample retrieval and end-to-end learning are essential for performance gains. Our framework demonstrates that query-specific adaptation significantly enhances both accuracy and generalization in PLA prediction.
Related
- RAGCell: retrieval-augmented generation as supervision for versatile single-cell analysis
- Patient classification using single-cell RNA-sequencing data with deep learning convolutional neural networks
- A Drug-Referenced Machine-Learning Framework for Prioritizing Korean Herbal Formulations in Hypertension Using Linear SVM and Graph Convolutional Networks
- Systematic comparison of affinity plasma proteomic technologies across multiple cohorts reveals generalizability of proteomic prediction
- Chemical and geometric representation fidelity improves drug--target affinity prediction
BibTeX
@article{abbasi2026rapla,
title = {{RA-PLA: Retrieval-augmented graph convolutional networks for protein-ligand binding affinity prediction}},
author = {Karim Abbasi and Hossein Banadkuki and Reza Sepahvand and Shima Rashidi},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0357278},
url = {https://doi.org/10.1371/journal.pone.0357278},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - RA-PLA: Retrieval-augmented graph convolutional networks for protein-ligand binding affinity prediction
AU - Karim Abbasi
AU - Hossein Banadkuki
AU - Reza Sepahvand
AU - Shima Rashidi
PY - 2026
DA - 2026/09/22
JO - PLOS One
DO - 10.1371/journal.pone.0357278
UR - https://doi.org/10.1371/journal.pone.0357278
AB - Protein-ligand binding affinity (PLA) prediction is essential for computational drug discovery, yet existing deep learning methods typically employ a single globally trained model that fails to adapt to individual query samples. In this paper, we propose 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). For each query, we automatically construct a graph where nodes represent protein-ligand pairs and edges encode pairwise similarities. Our model jointly learns joint descriptors, graph topology, and the GCN predictor. During inference, we fine-tune the model per query using its retrieved hard neighbors.We evaluate on four benchmarks: PDBbind, Davis, KIBA, and BindingDB. Our method consistently outperforms state-of-the-art approaches. On Davis, we achieve a Concordance Index (CI) of 0.952 ± 0.005 (1.6% improvement over NerLTR-DTA) and AUPR of 0.815 ± 0.005 (7.6% improvement). On KIBA, we achieve CI of 0.916 ± 0.002 and AUPR of 0.871 ± 0.005, outperforming DeepCDA by 2.7% and 5.9%, respectively. On PDBbind, our CI of 0.894 ± 0.002 represents an 11.4% improvement over PLA-MoRe. On BindingDB, we achieve AUPR of 0.518 ± 0.009 (5.9% improvement over DeepCDA). In cold-target generalization across four unseen protein families, we achieve average AUPR improvements of 15.5%. Paired t-tests confirm statistical significance (p < 0.05), and ablation studies validate that both hard sample retrieval and end-to-end learning are essential for performance gains. Our framework demonstrates that query-specific adaptation significantly enhances both accuracy and generalization in PLA prediction.
SN - 1932-6203
ER -