文献类型：期刊文献

英文题名：Drug-Target Interaction Prediction Based on Multi-Similarity Fusion and Sparse Dual-Graph Regularized Matrix Factorization

作者：Lian, Majun[1];Du, Wenli[1,2];Wang, Xinjie[1,2];Yao, Qian[1]

机构：[1]East China Univ Sci & Technol, Key Lab Smart Mfg Energy Chem Proc, Minist Educ, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200237, Peoples R China

年份：2021

卷号：9

起止页码：99718

外文期刊名：IEEE ACCESS

收录：;EI(收录号：20213010674954);WOS:【SCI-EXPANDED(收录号:WOS:000675198000001)】；

基金：This work was supported in part by the National Natural Science Foundation of China through the Basic Science Center Program under Grant 61988101; and in part by the National Natural Science Foundation of China under Grant 61725301, Grant 61890930-3, and Grant 61925305.

语种：英文

外文关键词：Drugs; Sparse matrices; Diseases; Feature extraction; Data models; Computational modeling; Predictive models; Drug-target interactions prediction; multi-similarity fusion; dual-graph regularized matrix factorization; manifold learning

摘要：Drug-target interactions (DTIs) prediction plays a vital role in drug discovery and design. Current studies typically use only standard drug similarity and target similarity, but the influence of known interactions has not been taken into account. In this paper, we propose an ensembled computational approach called multi-similarity fusion and sparse dual-graph regularized matrix factorization (MSDGRMF) for DTIs prediction. Specifically, different similarities are integrated to mine more useful information from the known interactions. The dual-graph regularized matrix factorization is used to predict the DTIs, in which the manifold learning is used for the low-dimensional representation of the drugs and targets data. In addition, not all the information of drug pairs and target pairs is useful. Thus, the useless information is discarded by sparse process. The proposed MSDGRMF is evaluated and compared on some benchmark datasets. Comparison results show that the MSDGRMF is better than some state-of-the-art approaches. More importantly, the proposed method can contribute to predicting potential DTIs.