Human Microbe-Disease Association Prediction Based on Adaptive Boosting

Area of Science:

• Microbiology
• Computational Biology
• Bioinformatics

Background:

• The human body hosts numerous microbes crucial for physiological processes.
• Emerging evidence links these microbes to various human diseases.
• Understanding microbe-disease associations is vital for advancing disease diagnosis and treatment.

Purpose of the Study:

• To develop and validate a computational model for predicting human microbe-disease associations.
• To identify novel microbe-disease relationships using machine learning.
• To provide a tool applicable to new diseases with unknown microbial links.

Main Methods:

• Development of an Adaptive Boosting for Human Microbe-Disease Association prediction (ABHMDA) model.
• Utilizing a strong classifier to calculate disease-microbe pair relation probabilities.
• Performance evaluation through global and local leave-one-out cross-validation (LOOCV).

Main Results:

• The ABHMDA model achieved high prediction accuracy with global LOOCV of 0.8869 and local LOOCV of 0.7910.
• Top predicted microbes for Asthma, Colorectal carcinoma, and Type 1 diabetes showed significant validation in literature and databases.
• The model demonstrated superior predictive performance in identifying relevant microbe-disease associations.

Conclusions:

• The ABHMDA model effectively predicts microbe-disease associations.
• This approach offers a promising strategy for discovering disease-related microbes.
• The model's accuracy supports its utility in disease research and potential clinical applications.