Related Experiment Video
Updated: Aug 23, 2025
06:16
mirMachine: A One-Stop Shop for Plant miRNA Annotation
Published on: May 1, 2021
2.6K
MEAHNE: miRNA-Disease Association Prediction Based on Semantic Information in a Heterogeneous Network
Chen Huang1, Keliang Cen1, Yang Zhang2
1School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), Shenzhen 518055, China.
Life (Basel, Switzerland)
|October 27, 2022
Summary
Predicting microRNA-disease associations is crucial for biomedical research. The proposed MEAHNE model effectively utilizes complex multisource data within a heterogeneous network, improving prediction accuracy for miRNA-disease pairs.
Area of Science:
- Bioinformatics
- Computational Biology
- Genomics
Background:
- Accurate prediction of microRNA-disease associations accelerates biomedical research.
- Existing methods struggle to leverage complex multisource data effectively, limiting prediction model performance.
Purpose of the Study:
- To propose a novel heterogeneous network prediction model (MEAHNE) for enhanced miRNA-disease association prediction.
- To fully exploit complex information within multisource data for improved predictive accuracy.
Main Methods:
- Constructed a heterogeneous network using collected multisource data.
- Developed the MEAHNE framework utilizing metapath instance semantic information and an attention mechanism.
- Employed end-to-end training for parameter optimization.
Main Results:
- MEAHNE demonstrated superior performance compared to state-of-the-art heterogeneous graph neural network methods, as evidenced by AUC and AUPRC metrics.
- The model successfully predicted 20 potential miRNA associations for breast cancer and 20 for nasopharyngeal cancer.
- Experimental validation confirmed 18 breast cancer-related and 14 nasopharyngeal cancer-related miRNAs.
Conclusions:
- MEAHNE effectively integrates multisource data within a heterogeneous network for accurate miRNA-disease association prediction.
- The proposed model offers a significant advancement in accelerating biomedical research through improved predictive capabilities.
