Related Experiment Video
Updated: Jun 21, 2025
07:35
A Knowledge Graph Approach to Elucidate the Role of Organellar Pathways in Disease via Biomedical Reports
Published on: October 13, 2023
1.6K
M3HOGAT: A Multi-View Multi-Modal Multi-Scale High-Order Graph Attention Network for Microbe-Disease Association
IEEE Journal of Biomedical and Health Informatics
|July 16, 2024
Summary
A new computational model, M³HOGAT, identifies microbes linked to human diseases by analyzing multi-source data. This method offers a faster, more cost-effective approach to understanding microbe-disease associations, aiding in disease research.
Area of Science:
- Microbiology
- Computational Biology
- Bioinformatics
Background:
- Growing evidence links human microbiome diversity to complex diseases.
- Traditional experimental methods for identifying microbe-disease associations are costly and time-consuming.
- Computational approaches are essential for efficient prediction of microbe-disease relationships.
Purpose of the Study:
- To propose a novel computational model, M³HOGAT, for predicting microbe-disease associations.
- To leverage multi-view, multi-modal network integration and multi-scale feature fusion for enhanced prediction accuracy.
- To provide a more efficient and cost-effective alternative to traditional experimental methods.
Main Methods:
- Construction of a microbe-disease association network and multiple similarity views using multi-source information.
- Development of a Higher-Order Graph Attention Network (HOGAT) to aggregate neighbor information from disparate orders for feature extraction.
- Implementation of a multi-scale feature fusion mechanism to learn interaction information from different views.
- Utilizing an inner product decoder to reconstruct the microbe-disease association matrix.
Main Results:
- M³HOGAT demonstrated superior performance compared to five state-of-the-art methods on the HMDAD and Disbiome datasets, confirmed by 5-fold cross-validation.
- Case studies on asthma and obesity validated the model's effectiveness in identifying potential disease-related microbes.
- The multi-view, multi-modal network and multi-scale feature fusion approach significantly improved prediction accuracy.
Conclusions:
- M³HOGAT provides an effective computational framework for predicting microbe-disease associations.
- The model's ability to integrate diverse data sources and features enhances the understanding of the human microbiome's role in disease.
- This approach has significant implications for future research in microbiome-based diagnostics and therapeutics.
