AUTHOR=Ma Yuanyuan , Liu Lifang , Chen Qianjun , Ma Yingjun 

TITLE=An Inductive Logistic Matrix Factorization Model for Predicting Drug-Metabolite Association With Vicus Regularization

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.650366

DOI=10.3389/fmicb.2021.650366

ISSN=1664-302X

ABSTRACT=Metabolites are closely related to human disease. The interaction between metabolites and drugs has drawn increasingly attentions in pharmacomicrobiomics. However, only a small portion of the drug-metabolite interactions were experimentally observed due to the fact that the experimental validation is costly, labor-intensive and time-consuming. Although a few computational approaches have been proposed to predict latent associations for various bipartite networks, such as miRNA-disease, drug-target interaction networks and so on, to our best knowledge the associations between drugs and metabolites have not been reported on a large scale. In this study, we propose a novel drug-metabolite association prediction method, namely Inductive Logistic Matrix Factorization (ILMF) to predict the latent associations between drugs and metabolites. Specifically, the proposed ILMF integrates drug-drug interaction, metabolite-metabolite interaction and drug-metabolite interaction into this framework, to model the probability that a drug would interact with a metabolite by logistic matrix factorization. Moreover, we exploit inductive matrix completion to guide the learning of projection matrices U, V with the optimal low- dimensional feature matrices of drugs and metabolites: , which can be obtained ahead via fusing multiple data source. Thus,  and can be viewed as the drug-specific and metabolite-specific latent representations instead of the original practice in LMF. And furthermore, we compute Vicus spectral matrix that reveals the refined local geometrical structure of the original data to obtain better prediction accuracy. Extensive experiments were conducted on a manually curated ‘DrugMetaboliteAtlas’ dataset. The experimental results show that ILMF can achieve competitive performance compared with other state-of-the-art approaches, which demonstrate its effectiveness in predicting potential drug-metabolite associations.