Learning to Denoise Biomedical Knowledge Graph for Robust Molecular Interaction Prediction

TL;DR

This paper introduces BioKDN, a novel denoising network for biomedical knowledge graphs that improves molecular interaction predictions by reducing noise and enhancing semantic consistency, outperforming existing models.

Contribution

BioKDN is the first learnable framework that denoises noisy links and maintains semantic robustness in biomedical knowledge graphs for molecular interaction prediction.

Findings

BioKDN outperforms state-of-the-art models in DTI and DDI prediction tasks.

It effectively denoises unreliable interactions in contaminated knowledge graphs.

BioKDN enhances the robustness and reliability of molecular interaction predictions.

Abstract

Molecular interaction prediction plays a crucial role in forecasting unknown interactions between molecules, such as drug-target interaction (DTI) and drug-drug interaction (DDI), which are essential in the field of drug discovery and therapeutics. Although previous prediction methods have yielded promising results by leveraging the rich semantics and topological structure of biomedical knowledge graphs (KGs), they have primarily focused on enhancing predictive performance without addressing the presence of inevitable noise and inconsistent semantics. This limitation has hindered the advancement of KG-based prediction methods. To address this limitation, we propose BioKDN (Biomedical Knowledge Graph Denoising Network) for robust molecular interaction prediction. BioKDN refines the reliable structure of local subgraphs by denoising noisy links in a learnable manner, providing a general module for extracting task-relevant interactions. To enhance the reliability of the refined structure, BioKDN maintains consistent and robust semantics by smoothing relations around the target interaction. By maximizing the mutual information between reliable structure and smoothed relations, BioKDN emphasizes informative semantics to enable precise predictions. Experimental results on real-world datasets show that BioKDN surpasses state-of-the-art models in DTI and DDI prediction tasks, confirming the effectiveness and robustness of BioKDN in denoising unreliable interactions within contaminated KGs