Predicting and understanding comprehensive drug-drug interactions via semi-nonnegative matrix factorization

Hui Yu,Kai Dong,Jian-Yu Shi,Hua Huang,Siu-Ming Yiu,Kui-Tao Mao,Zhi Chen

doi:10.1186/s12918-018-0532-7

Abstract

BackgroundDrug-drug interactions (DDIs) always cause unexpected and even adverse drug reactions. It is important to identify DDIs before drugs are used in the market. However, preclinical identification of DDIs requires much money and time. Computational approaches have exhibited their abilities to predict potential DDIs on a large scale by utilizing pre-market drug properties (e.g. chemical structure). Nevertheless, none of them can predict two comprehensive types of DDIs, including enhancive and degressive DDIs, which increases and decreases the behaviors of the interacting drugs respectively. There is a lack of systematic analysis on the structural relationship among known DDIs. Revealing such a relationship is very important, because it is able to help understand how DDIs occur. Both the prediction of comprehensive DDIs and the discovery of structural relationship among them play an important guidance when making a co-prescription.ResultsIn this work, treating a set of comprehensive DDIs as a signed network, we design a novel model (DDINMF) for the prediction of enhancive and degressive DDIs based on semi-nonnegative matrix factorization. Inspiringly, DDINMF achieves the conventional DDI prediction (AUROC = 0.872 and AUPR = 0.605) and the comprehensive DDI prediction (AUROC = 0.796 and AUPR = 0.579). Compared with two state-of-the-art approaches, DDINMF shows it superiority. Finally, representing DDIs as a binary network and a signed network respectively, an analysis based on NMF reveals crucial knowledge hidden among DDIs.ConclusionsOur approach is able to predict not only conventional binary DDIs but also comprehensive DDIs. More importantly, it reveals several key points about the DDI network: (1) both binary and signed networks show fairly clear clusters, in which both drug degree and the difference between positive degree and negative degree show significant distribution; (2) the drugs having large degrees tend to have a larger difference between positive degree and negative degree; (3) though the binary DDI network contains no information about enhancive and degressive DDIs at all, it implies some of their relationship in the comprehensive DDI matrix; (4) the occurrence of signs indicating enhancive and degressive DDIs is not random because the comprehensive DDI network is equipped with a structural balance.

Highlights

Drug-drug interactions (DDIs) always cause unexpected and even adverse drug reactions
Existing computational approaches are able to screen potential DDIs on a large scale before drugs are used in the medicine market
None of them can predict comprehensive DDIs, including enhancive and degressive DDIs, though it is important to know whether the interaction increases or decreases the behavior of the interacting drugs before making a co-prescription

Summary

Introduction

Drug-drug interactions (DDIs) always cause unexpected and even adverse drug reactions. When two or more drugs are taken together, they would unexpectedly influence each other in terms of pharmacokinetic or pharmacodynamical behavior [1] This kind of influence is termed as Drug-Drug Interaction (DDI), which would cause adverse drug reactions (e.g. the reduction in efficacy or the increment on unexpected toxicity among the co-prescribed drugs). As the number of approved drugs increases, the number of unidentified DDIs is rapidly rising, such that more and more adverse effects among the drugs may occur. Traditional experimental approaches for DDI identification (e.g. testing cytochrome P450 [7] or transporter-associated interactions [8]) face challenges, such as high cost, long duration, animal welfare considerations [9], the very limited number of participants and the great number of drug combinations under screening in clinical trials. Only a few of DDIs could be identified during drug development (usually the clinical trial phase), and most of them are reported after the drugs are approved

Methods

Results

Conclusion