Alterations in Gut Microbiota-Brain Axis in Major Depressive Disorder as Identified by Machine Learning

Area of Science:

• Microbiome research
• Gut-brain axis
• Artificial intelligence in health

Background:

• Major depressive disorder (MDD) is complex, influenced by biological, ecological, and social factors.
• The gut-brain axis is a key area for understanding MDD, linking gut microbiota, diet, and environment.
• Statistical and artificial intelligence (AI) tools are advancing research in this field.

Purpose of the Study:

• To identify potential gut microbial biomarkers for MDD using AI.
• To analyze data from the American Gut Project (AGP) for differences in gut microbiota between MDD patients and controls.
• To explore the utility of machine learning (ML) in biomarker discovery for MDD.

Main Methods:

• Analysis of 361 control and 23 MDD samples from the American Gut Project (AGP).
• Application of differential abundance tools and machine learning models (Random Forest, XGBoost).
• Utilized Shapley Additive Explanations (SHAP) values for interpreting ML model results.

Main Results:

• No significant differences in alpha and beta diversity were found between groups, except for age.
• Machine learning models consistently identified a decrease in Bifidobacterium adolescentis in MDD samples.
• Increases in Odoribacter, Ruminococcus, and Adlercreutzia were consistently detected in MDD samples.

Conclusions:

• Bifidobacterium adolescentis, Odoribacter, Ruminococcus, and Adlercreutzia show potential as biomarkers for MDD.
• These microbes influence inflammation, neurotransmitter balance, and gut permeability, pathways relevant to depression.
• ML shows promise for decoding the gut-brain axis, aiding biomarker discovery and personalized MDD treatment.