🤖 AI Summary
Existing deep graph models struggle to accurately predict signed chemical–gene interactions (e.g., activation/inhibition) in biological networks and fail to distinguish polarity-specific relational semantics. Method: We propose a novel signed graph representation learning framework that integrates graph convolutional networks (GCNs) with high-order tensor decomposition—marking the first application of tensor decomposition to signed graph embedding, explicitly modeling interaction polarity and overcoming GCNs’ inherent limitations in capturing signed relational semantics. Contribution/Results: Evaluated on multiple real-world biological network datasets, our model achieves significant gains in prediction accuracy—improving AUC by over 8% under sparse and noisy conditions. Moreover, it successfully elucidates mechanistic pathways for several known drugs, demonstrating both practical utility and interpretability in mechanism-driven drug repurposing.
📝 Abstract
In pharmaceutical research, the strategy of drug repurposing accelerates the development of new therapies while reducing R&D costs. Network pharmacology lays the theoretical groundwork for identifying new drug indications, and deep graph models have become essential for their precision in mapping complex biological networks. Our study introduces an advanced graph model that utilizes graph convolutional networks and tensor decomposition to effectively predict signed chemical-gene interactions. This model demonstrates superior predictive performance, especially in handling the polar relations in biological networks. Our research opens new avenues for drug discovery and repurposing, especially in understanding the mechanism of actions of drugs.