1st International and 10th National Iranian Conference on Bioinformatics
Home Accepted Papers
Indicator Regularized Non-Negative Matrix Factorization Vs. a Novel Combinatorial Heuristic Matrix Factorization: a Comparison of Matrix Factorization Methods as the Building Block of Drug Repurposing
Paper ID : 1234-ICB10
Authors:
Mohsen Hooshmand *1, Arash Zabihian2, Sajjad Gharaghani3
1Department of Computer Science and Information Technology,Institute for Advanced Studies in Basic Sciences, Zanjan, Iran
2Laboratory of Bioinformatics and Drug Design (LBD), Institute of Biochemistry and Biophysics, University of Tehran, Tehran, Iran
3Laboratory of Bioinformatics and Drug Design (LBD), Institute of Biochemistry and Biophysics, University of Tehran, Tehran, Iran.
Abstract:
Drug repurposing is one of the ponderable computational methods suggested for rapid developing of a drug. This approach can be formulated as a matrix factorization (MF) method. First, a MF method is applied to different types of drug-target relation and similarity matrices. Then, the unseen data will be analysed to evalutae the predictive power of the model. Some types of data in bioinformatics are binary, e.g. binary representation of impacts of antiviral drugs on viruses. MF of such data is more complicated than the conventional numerical values. Most of conventional algorithms for binary matrix factorization use gradient-based methods. These methods utilize relaxation approach to solve continuous binary problems. The relaxation approach turns the binary constraint into a box constraint [1]. Despite the widespread use, these methods do not perform well on sparse data sets, and also unable to return a proper approximation of problem. To avoid facing these limitations, we propose a binary matrix factorization method utilizing combinatorial optimization and modular arithmetic. In our prposal, binary values are used in each step and the results come from modular multiplication. We compare our results with Indicator Regularized Non-Negative Matrix Factorization (IRNMF) which is a gradient based method [2]. Both methods are appled to the same antiviral-viruse interaction matrix and Five-fold cross-validation (CV) are used to report the performance. The results indicate that better performance and lower error value are the advantages of the suggested method in compare with IRNMF.
Keywords:
Matrix Factorization; Drug repurposing; Antiviral Drugs; Heuristic Factorization
Status : Paper Accepted (Oral Presentation)