Related Experiment Video
Updated: Aug 28, 2025
Author Spotlight: Integrated Multi-Omics Analysis for Unveiling Multicellular Immune Signatures in Clinical Heart Attack Cohorts
Published on: September 20, 2024
MHDMF: Prediction of miRNA-disease associations based on Deep Matrix Factorization with Multi-source Graph
Ning Ai1, Yong Liang2, Hao-Laing Yuan3
1Peng Cheng Laboratory, Shenzhen 518005, China; School of Computer Science and Engineering, Faculty of Innovation Engineering, Macau University of Science and Technology, Avenida Wai Long, Taipa, China.
This study introduces MHDMF, a computational framework that improves the prediction of disease-microRNA associations by correcting false negatives using multi-source information and graph convolutional networks. The method enhances accuracy in identifying crucial biomarkers for various diseases.
Area of Science:
- Biomedical Informatics
- Computational Biology
- Genomics
Background:
- MicroRNAs (miRNAs) are vital biomarkers in disease processes.
- Computational methods complement costly experimental approaches for miRNA-disease association prediction.
- Existing methods struggle with high false-negative rates and integrating multi-source data.
Purpose of the Study:
- To develop an end-to-end computational framework (MHDMF) for discovering latent disease-miRNA associations.
- To address challenges in utilizing high false-negative associations and multi-source information.
- To improve the accuracy and efficiency of miRNA-disease association prediction.
Main Methods:
- Developed MHDMF, integrating multi-source information on a heterogeneous network.
- Utilized multi-source Graph Convolutional Network (GCN) to correct false-negative miRNA-disease associations.
- Employed Deep Matrix Factorization (DMF) on a reformulated score matrix for prediction.
Main Results:
- MHDMF significantly outperforms existing methods in miRNA-disease association prediction.
- Experimental results demonstrate the framework's superior performance.
- Case studies indicate MHDMF's convenience and efficiency.
Conclusions:
- MHDMF offers an effective computational approach for predicting disease-miRNA associations.
- The framework provides a novel perspective for miRNA-disease association discovery.
- MHDMF shows potential as a valuable tool in biomedical research.
