AUTHOR=Guan Yong-Jian , Yu Chang-Qing , Li Li-Ping , You Zhu-Hong , Ren Zhong-Hao , Pan Jie , Li Yue-Chao 

TITLE=BNEMDI: A Novel MicroRNA–Drug Interaction Prediction Model Based on Multi-Source Information With a Large-Scale Biological Network

JOURNAL=Frontiers in Genetics

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.919264

DOI=10.3389/fgene.2022.919264

ISSN=1664-8021

ABSTRACT=MicroRNA (miRNA) can regulate gene expression to produce specific proteins and it is a new target in pharmacy. To date, there are many experiments leverage miRNA to reveal drug efficacy and pathogenesis at the molecular level. As we all know that conventional wet experiments suffer from many problems, including time-consuming, labor-intensive and costly. Thus, there is an urgent need to develop a novel computational model to facilitate the identification of miRNA-drug interactions (MDIs). In this work, we proposed a novel bipartite network embedding-based method called BNEMDI to predict MDIs. Firstly, the Bipartite Network Embedding (BiNE) algorithm was employed to learn the topological features from the network. Then, the inherent attribute of drugs and miRNAs were expressed as attribute features by MACCS fingerprints and k-mers. Finally, we fed both features into Deep Neural Network (DNN) for integrating and training. To validate the prediction ability of BNEMDI model, we applied it on five different MDI datasets under 5-fold cross-validation. And the proposed model obtained excellent AUC values of 0.9568, 0.9420, 0.8489, 0.8774 and 0.9005 in ncDR, RNAInter, SM2miR1, SM2miR2 and SM2miR MDIs datasets, respectively. To further verify the performance of BNEMDI, we compared it with some existing powerful methods. We also compared BiNE with several different network embedding methods. Furthermore, we carried out a case study on a common drug, which named 5- fluorouracil. Among the top 50 miRNAs predicted by BNEMDI, there were 38 verified by experimental literature. The comprehensive experiment results demonstrated that BNEMDI is effective and robust for predicting MDIs. In future work, we hope the proposed model can be a reliable supplement method for the development of pharmacology and miRNA therapeutics.