Predicting drug-drug interactions: A deep learning approach with GCN-based collaborative filtering.

Artif Intell Med

Oncocross Co. Ltd., #905, C Block, Beobwon-ro 11-gil, Songpa-gu, Seoul, 05836, Republic of Korea; Department of Regulatory Science, Kyung Hee University, 26, Kyungheedae-ro, Dongdaemun-gu, Seoul, 02447, Republic of Korea. Electronic address:

Published: September 2025


Category Ranking

98%

Total Visits

921

Avg Visit Duration

2 minutes

Citations

20

Article Abstract

The use of combination drugs among patients is increasing due to effectiveness compared to monotherapies. However, healthcare providers should continue to be concerned about the potential risks associated with patient safety arising from drug-drug interactions (DDIs) when they use combination drugs. Whereas direct physicochemical interactions contribute to certain cases of DDIs, the majority of DDIs occur because one drug modulates enzymes, such as cytochrome P450, responsible for metabolizing another drug. Therefore, drugs that interact with the same family drugs are more likely to interact with each other by mediating specific enzymes. Adapted from techniques used to recommend users with similar interests, we introduce an AI recommendation model with graph convolutional network (GCN) and collaborative filtering that analyzes the connectivity of interacting drugs rather than their chemical structures. This approach deviates from typical classification models by not requiring sampling of undefined interactions as negative samples, allowing the prediction of potential interactions for all unknown drug pairs, circumventing the challenges associated with selecting negative interactions and data imbalance. Our methodology used the DrugBank database (version 5.1.9 released on January 3, 2022), encompassing 4,072 drugs and 1,391,790 drug pairs with interactions. Furthermore, the robustness of the model was verified through a 5-fold validation and external data validation using TWOSIDES data. Notably, our model's efficacy is established solely through the exploitation of DDI reports, offering a versatile framework capable of accurately predicting interactions among diverse drug types. The source code for this project is distributed on GitHub (https://github.com/yeonuk-Jeong/DDI-OCF).

Download full-text PDF

Source
http://dx.doi.org/10.1016/j.artmed.2025.103185DOI Listing

Publication Analysis

Top Keywords

interactions
8
drug-drug interactions
8
collaborative filtering
8
combination drugs
8
drugs interact
8
drug pairs
8
drugs
6
drug
5
predicting drug-drug
4
interactions deep
4

Similar Publications

3-O-sulfation of heparan sulfate (HS) is the key determinant for binding and activation of Antithrombin III (AT). This interaction is the basis of heparin treatment to prevent thrombotic events and excess coagulation. Antithrombin-binding HS (HSAT) is expressed in human tissues, but is thought to be expressed in the subendothelial space, mast cells, and follicular fluid.

View Article and Find Full Text PDF

Sorting nexin 3 promotes ischemic retinopathy through RIP1- and RIP3-mediated myeloid cell necroptosis and mitochondrial fission.

Proc Natl Acad Sci U S A

September 2025

State Key Laboratory of Bioactive Molecules and Druggability Assessment, Guangdong Province Key Laboratory of Pharmacodynamic Constituents of Traditional Chinese Medicine and New Drugs Research, International Cooperative Laboratory of Traditional Chinese Medicine Modernization and Innovative Drug De

Proliferative retinopathy is a leading cause of irreversible blindness in humans; however, the molecular mechanisms behind the immune cell-mediated retinal angiogenesis remain poorly elucidated. Here, using single-cell RNA sequencing in an oxygen-induced retinopathy (OIR) model, we identified an enrichment of sorting nexin (SNX)-related pathways, with SNX3, a member of the SNX family that is involved in endosomal sorting and trafficking, being significantly upregulated in the myeloid cell subpopulations of OIR retinas. Immunostaining showed that SNX3 expression is markedly increased in the retinal microglia/macrophages of mice with OIR, which is mainly located within and around the neovascular tufts.

View Article and Find Full Text PDF

Retinal ganglion cells (RGCs) are highly compartmentalized neurons whose long axons serve as the sole connection between the eye and the brain. In both injury and disease, RGC degeneration occurs in a similarly compartmentalized manner, with distinct molecular and cellular responses in the axonal and somatodendritic regions. The goal of this study was to establish a microfluidic-based platform to investigate RGC compartmentalization in both health and disease states.

View Article and Find Full Text PDF

Agonist-induced interaction of G protein-coupled receptors (GPCRs) with β-arrestins (βarrs) is a critical mechanism that regulates the spatiotemporal pattern of receptor localization and signaling. While the underlying mechanism governing GPCR-βarr interaction is primarily conserved and involves receptor activation and phosphorylation, there are several examples of receptor-specific fine-tuning of βarr-mediated functional outcomes. Considering the key contribution of conformational plasticity of βarrs in driving receptor-specific functional responses, it is important to develop novel sensors capable of reporting distinct βarr conformations in cellular context.

View Article and Find Full Text PDF

Programmable self-assembly has recently enabled the creation of complex structures through precise control of the interparticle interactions and the particle geometries. Targeting ever more structurally complex, dynamic, and functional assemblies necessitates going beyond the design of the structure itself, to the measurement and control of the local flexibility of the intersubunit connections and its impact on the collective mechanics of the entire assembly. In this study, we demonstrate a method to infer the mechanical properties of multisubunit assemblies using cryogenic electron microscopy (cryo-EM) and RELION's multi-body refinement.

View Article and Find Full Text PDF