SPARSE: a sparse hypergraph neural network for learning multiple types of latent combinations to accurately predict drug–drug interactions
Open Access
- 27 June 2022
- journal article
- research article
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 38 (Supplement), i333-i341
- https://doi.org/10.1093/bioinformatics/btac250
Abstract
Predicting side effects of drug–drug interactions (DDIs) is an important task in pharmacology. The state-of-the-art methods for DDI prediction use hypergraph neural networks to learn latent representations of drugs and side effects to express high-order relationships among two interacting drugs and a side effect. The idea of these methods is that each side effect is caused by a unique combination of latent features of the corresponding interacting drugs. However, in reality, a side effect might have multiple, different mechanisms that cannot be represented by a single combination of latent features of drugs. Moreover, DDI data are sparse, suggesting that using a sparsity regularization would help to learn better latent representations to improve prediction performances. We propose SPARSE, which encodes the DDI hypergraph and drug features to latent spaces to learn multiple types of combinations of latent features of drugs and side effects, controlling the model sparsity by a sparse prior. Our extensive experiments using both synthetic and three real-world DDI datasets showed the clear predictive performance advantage of SPARSE over cutting-edge competing methods. Also, latent feature analysis over unknown top predictions by SPARSE demonstrated the interpretability advantage contributed by the model sparsity. Code and data can be accessed at https://github.com/anhnda/SPARSE. Supplementary data are available at Bioinformatics online.Funding Information
- Otsuka Toshimi Scholarship Foundation
- MEXT KAKENHI (22K12150)
- Japan Society for the Promotion of Science
- Postdoctoral Fellowships for Research in Japan (P20809, MEXT KAKENHI, 19H04169, 20F20809, 21H05027, 22H03645)
- Academy of Finland
This publication has 24 references indexed in Scilit:
- Is medical research informing professional practice more highly cited? Evidence from AHFS DI Essentials in drugs.comScientometrics, 2017
- Sparsity information and regularization in the horseshoe and other shrinkage priorsElectronic Journal of Statistics, 2017
- The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced DatasetsPLOS ONE, 2015
- Cytochrome P450-Mediated Bioactivation of Mefenamic Acid to Quinoneimine Intermediates and Inactivation by Human Glutathione S-TransferasesChemical Research in Toxicology, 2014
- Data-Driven Prediction of Drug Effects and InteractionsScience Translational Medicine, 2012
- Epidemiology and characteristics of adverse drug reactions caused by drug–drug interactionsExpert Opinion on Drug Safety, 2011
- Different Mechanisms in Formation and Prevention of Indomethacin-induced Gastric UlcersInflammation, 2010
- Cardiovascular Tolerability and Safety of Triptans: A Review of Clinical DataHeadache: The Journal of Head and Face Pain, 2004
- Specific GABA A Circuits for Visual Cortical PlasticityScience, 2004
- Barbiturate-like actions of the propanediol dicarbamates felbamate and meprobamate.1997