Yiwen Sun1, Zexuan Zhu2, Zhu-Hong You3, Zijie Zeng2, Zhi-An Huang4, Yu-An Huang5. 1. School of Medicine, Shenzhen University, Shenzhen, 518060, China. 2. College of Computer Science and Software Engineering, Shenzhen University, Shenzhen, 518060, China. 3. Xinjiang Technical Institute of Physics and Chemistry, Chinese Academy of Science, ürümqi, 830011, China. 4. Department of Computer Science, City University of Hong Kong, Hong Kong, 999077, China. huang_za@outlook.com. 5. Department of Computing, Hong Kong Polytechnic University, Hong Kong, 999077, China. yahaung1991@gmail.com.
Abstract
BACKGROUND: MicroRNA (miRNA) plays a key role in regulation mechanism of human biological processes, including the development of disease and disorder. It is necessary to identify potential miRNA biomarkers for various human diseases. Computational prediction model is expected to accelerate the process of identification. RESULTS: Considering the limitations of previously proposed models, we present a novel computational model called FMSM. It infers latent miRNA biomarkers involved in the mechanism of various diseases based on the known miRNA-disease association network, miRNA expression similarity, disease semantic similarity and Gaussian interaction profile kernel similarity. FMSM achieves reliable prediction performance in 5-fold and leave-one-out cross validations with area under ROC curve (AUC) values of 0.9629+/- 0.0127 and 0.9433, respectively, which outperforms the state-of-the-art competitors and classical algorithms. In addition, 19 of top 25 predicted miRNAs have been validated to have associations with Colonic Neoplasms in case study. CONCLUSIONS: A factored miRNA similarity based model and miRNA expression similarity substantially contribute to the well-performing prediction. The list of the predicted most latent miRNA biomarkers of various human diseases is publicized. It is anticipated that FMSM could serve as a useful tool guiding the future experimental validation for those promising miRNA biomarker candidates.
BACKGROUND: MicroRNA (miRNA) plays a key role in regulation mechanism of human biological processes, including the development of disease and disorder. It is necessary to identify potential miRNA biomarkers for various human diseases. Computational prediction model is expected to accelerate the process of identification. RESULTS: Considering the limitations of previously proposed models, we present a novel computational model called FMSM. It infers latent miRNA biomarkers involved in the mechanism of various diseases based on the known miRNA-disease association network, miRNA expression similarity, disease semantic similarity and Gaussian interaction profile kernel similarity. FMSM achieves reliable prediction performance in 5-fold and leave-one-out cross validations with area under ROC curve (AUC) values of 0.9629+/- 0.0127 and 0.9433, respectively, which outperforms the state-of-the-art competitors and classical algorithms. In addition, 19 of top 25 predicted miRNAs have been validated to have associations with Colonic Neoplasms in case study. CONCLUSIONS: A factored miRNA similarity based model and miRNA expression similarity substantially contribute to the well-performing prediction. The list of the predicted most latent miRNA biomarkers of various human diseases is publicized. It is anticipated that FMSM could serve as a useful tool guiding the future experimental validation for those promising miRNA biomarker candidates.
Entities:
Keywords:
Biomarker; Computational prediction; Expression profiles; miRNA-disease association
Authors: B J Reinhart; F J Slack; M Basson; A E Pasquinelli; J C Bettinger; A E Rougvie; H R Horvitz; G Ruvkun Journal: Nature Date: 2000-02-24 Impact factor: 49.962