Recent Developments in GNNs for Drug Discovery

2025-06-02

AI Summary

Existing surveys on graph neural networks (GNNs) in computational drug discovery lack systematic, structured taxonomies—particularly across input representations and downstream task paradigms—hindering comparative analysis and methodological advancement. Method: This work presents a comprehensive review of recent GNN advances for molecular generation, property prediction, and drug–drug interaction prediction. We introduce a novel dual-axis classification framework: (i) by input representation type (e.g., atom/bond-level, subgraph-level, 3D geometric), and (ii) by downstream task paradigm (e.g., single- vs. multi-task, supervised vs. self-supervised). Benchmark evaluations are unified across QM9, MoleculeNet, DrugBank, and other standard datasets. Contribution/Results: We establish an end-to-end methodology framework spanning representation learning, model design, and evaluation. Key insights include fundamental capability limits (e.g., modeling long-range dependencies, generalizing across 3D conformers) and persistent challenges (e.g., limited interpretability). The review delivers a structured, knowledge-graph–inspired synthesis—serving as a reproducible, cross-disciplinary guide for algorithmic innovation and translational application.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Natural Language Processing: Generation

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

In this paper, we review recent developments and the role of Graph Neural Networks (GNNs) in computational drug discovery, including molecule generation, molecular property prediction, and drug-drug interaction prediction. By summarizing the most recent developments in this area, we underscore the capabilities of GNNs to comprehend intricate molecular patterns, while exploring both their current and prospective applications. We initiate our discussion by examining various molecular representations, followed by detailed discussions and categorization of existing GNN models based on their input types and downstream application tasks. We also collect a list of commonly used benchmark datasets for a variety of applications. We conclude the paper with brief discussions and summarize common trends in this important research area.

Problem

• Review GNNs' role in computational drug discovery
• Explore GNNs for molecule and property prediction
• Summarize GNN models and benchmark datasets

Innovation

• GNNs analyze complex molecular patterns
• GNNs classify models by input types
• GNNs utilize benchmark datasets widely