Graph Diffusion Network for Drug-Gene Prediction

TL;DR

This paper introduces GDNDGP, a graph diffusion network that improves drug-gene prediction by capturing complex relationships and generating hard negatives efficiently, outperforming existing methods on key datasets.

Contribution

The paper presents a novel graph diffusion network with meta-path-based learning and parallel diffusion for hard negative sampling, addressing data sparsity and efficiency issues in drug-gene prediction.

Findings

Achieves superior performance on DGIdb 4.0 dataset

Demonstrates strong generalization on drug-gene-disease networks

Significantly outperforms existing methods in accuracy

Abstract

Predicting drug-gene associations is crucial for drug development and disease treatment. While graph neural networks (GNN) have shown effectiveness in this task, they face challenges with data sparsity and efficient contrastive learning implementation. We introduce a graph diffusion network for drug-gene prediction (GDNDGP), a framework that addresses these limitations through two key innovations. First, it employs meta-path-based homogeneous graph learning to capture drug-drug and gene-gene relationships, ensuring similar entities share embedding spaces. Second, it incorporates a parallel diffusion network that generates hard negative samples during training, eliminating the need for exhaustive negative sample retrieval. Our model achieves superior performance on the DGIdb 4.0 dataset and demonstrates strong generalization capability on tripartite drug-gene-disease networks. Results show significant improvements over existing methods in drug-gene prediction tasks, particularly in handling complex heterogeneous relationships. The source code is publicly available at https://github.com/csjywu1/GDNDGP.