Related Experiment Video
Updated: Jun 16, 2025
09:47
Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
998
Identification of microbe-disease signed associations via multi-scale variational graph autoencoder based on signed
Huan Zhu1, Hongxia Hao2, Liang Yu3
1School of Computer Science and Technology, Xidian University, Xi'an, China.
BMC Biology
|August 15, 2024
Summary
This study introduces MSignVGAE, a novel framework for predicting microbe-disease sign associations. The method accurately identifies microbial links to diseases, aiding precision medicine and early diagnosis.
Area of Science:
- Microbiology
- Bioinformatics
- Computational Biology
Background:
- The human microbiome plays a critical role in health and disease.
- Identifying disease-associated microbes is vital for early diagnosis and precision medicine.
Purpose of the Study:
- To develop a framework for predicting microbe-disease sign associations.
- To enhance the understanding of microbial roles in disease states.
Main Methods:
- MSignVGAE framework utilizes signed message propagation and a graph variational autoencoder.
- Employs multi-scale concepts and denoising autoencoder for robust data representation.
- Models microbe-disease associations as a heterogeneous graph with similarity features.
- Utilizes XGBoost for predicting sign associations.
Main Results:
- MSignVGAE achieves high performance with AUROC of 0.9742 and AUPR of 0.9601.
- Demonstrates effective capture of association distributions in case studies.
- Successfully leverages signed information for accurate predictions.
Conclusions:
- MSignVGAE accurately predicts microbe-disease sign associations.
- The framework provides a comprehensive understanding of microbial roles in diseases.
- Highlights the potential of MSignVGAE in advancing precision medicine.
