Inferring Disease-Associated Microbes Based on Multi-Data Integration and Network Consistency Projection.

Area of Science:

• Microbiology
• Computational Biology
• Genomics

Background:

• The human microbiome plays a crucial role in cellular physiology and is increasingly linked to complex human diseases.
• Understanding microbe-disease associations is vital for developing novel therapeutic, diagnostic, and preventive strategies.
• Existing computational models for microbe-disease association prediction often suffer from ineffective fusion of similarity networks.

Purpose of the Study:

• To propose a novel computational model, HMDA-Pred, for predicting Human Microbe-Disease Associations (HMDA).
• To effectively fuse multiple microbe similarity networks using a linear network fusion method.
• To enhance the accuracy and reliability of microbe-disease association predictions.

Main Methods:

• Developed HMDA-Pred, a computational model integrating multi-data and employing network consistency projection.
• Utilized a linear network fusion method to combine diverse microbe similarity networks.
• Performed rigorous validation using leave-one-out cross-validation (LOOCV) and 5-fold cross-validation (5-fold CV).

Main Results:

• HMDA-Pred achieved high performance with AUC values of 0.9589 (LOOCV) and 0.9361 ± 0.0037 (5-fold CV).
• Case studies demonstrated significant predictive power, with top predictions validated in literature for asthma, colon cancer, and inflammatory bowel disease.

Conclusions:

• The proposed HMDA-Pred model effectively integrates multiple similarity networks for accurate microbe-disease association prediction.
• HMDA-Pred offers a promising computational tool for advancing the understanding of microbe-driven diseases.
• The model's validated predictions provide a foundation for future research in disease etiology and targeted interventions.