A Graph Convolutional Matrix Completion Method for miRNA-Disease Association Prediction

Autor:	Wei Wang, Cong Shen, Jiawei Luo, Nguyen Hoang Tu
Rok vydání:	2020
Předmět:	Matrix completion Computer science business.industry In silico Disease Association Disease Disease pathogenesis Machine learning computer.software_genre Graph (abstract data type) Artificial intelligence business computer Prior information
Zdroj:	Intelligent Computing Theories and Application ISBN: 9783030608019 ICIC (2)
Popis:	MicroRNAs (miRNAs) play a key role in various biological processes associated with human diseases. Identification of miRNA-disease relationships can help to understand disease pathogenesis. Experimentally verifying substantial associations between miRNAs and diseases is the most convincing but time-consuming, while in silico methods can provide efficient alternatives. However, existing computational methods still have room for improvement in considering topology and prior information of network nodes. In this paper, we presented a novel model called GCMCAP, in which we referred to the prediction of potential miRNA-disease associations as a recommendation problem. In our framework, we integrated graph convolution networks as feature extractors into a matrix completion to predict diseases related miRNAs. We tested GCMCAP and other three methods on the same dataset. The results indicate that GCMCAP outperforms other methods with respect to average AUC value. In addition, case studies show that GCMCAP has a great capability to discover novel miRNA-disease associations.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::77af28ecbe46595a9dc2956acc1d8ea4 https://doi.org/10.1007/978-3-030-60802-6_18 Zobrazit plný text záznamu