Prediction of multiple types of drug interactions based on multi-scale fusion and dual-view fusion

• DOI: 10.3389/fphar.2024.1354540
• Publication Date: 2024-02-16T04:59:43Z
• Authors: Dawei Pan; Ping Lu; Yunbing Wu; Liping Kang; Fengxin Huang; Kaibiao Lin; Fan Yang
• Keywords: 0301 basic medicine; Artificial intelligence; multi-scale fusion; Scale (ratio); Feature (linguistics); FOS: Political science; Biomedical Ontologies and Text Mining; Graph; Political science; drug drug interaction prediction; Psychiatry; Dual (grammatical number); Physics; Politics; Life Sciences; Advances in Metabolomics Research; FOS: Philosophy, ethics and religion; Computational Theory and Mathematics; Physical Sciences; Medicine; Drug; Art; Computational Methods in Drug Discovery; Drug Target Identification; graph neural network; Convolutional neural network; FOS: Law; RM1-950; Quantum mechanics; 03 medical and health sciences; Theoretical computer science; Biochemistry, Genetics and Molecular Biology; Machine learning; graph features represent learning; Fusion; Molecular Biology; Data mining; Pharmacology; Feature learning; Linguistics; multi-class classification; Computer science; Philosophy; Literature; Computer Science; FOS: Languages and literature; Therapeutics. Pharmacology; Representation (politics); Law

Potential drug-drug interactions (DDI) can lead to adverse drug reactions (ADR), and DDI prediction can help pharmacy researchers detect harmful DDI early. However, existing DDI prediction methods fall short in fully capturing drug information. They typically employ a single-view input, focusing solely on drug features or drug networks. Moreover, they rely exclusively on the final model layer for predictions, overlooking the nuanced information present across various network layers. To address these limitations, we propose a multi-scale dual-view fusion (MSDF) method for DDI prediction. More specifically, MSDF first constructs two views, topological and feature views of drugs, as model inputs. Then a graph convolutional neural network is used to extract the feature representations from each view. On top of that, a multi-scale fusion module integrates information across different graph convolutional layers to create comprehensive drug embeddings. The embeddings from the two views are summed as the final representation for classification. Experiments on two real-world datasets demonstrate that MSDF achieves higher accuracy than state-of-the-art methods, as the dual-view, multi-scale approach better captures drug characteristics.