Dietary and Gut Microbiome Features Associated with Depression Status: An Exploratory Machine-Learning Study
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: The gut microbiome contributes to depression through the microbiota–gut–brain axis, involving immune, metabolic, neuroendocrine, and neural pathways. Diet is an important determinant of microbial composition and may therefore modulate depression-related pathways. However, despite of the growing evidence connecting diet, the microbiome, and depressive symptoms, the underlying mechanistic insights remain limited, in part because of the challenges in capturing these complex and potentially non-linear relationships. As an alternative to conventional mediation analyses, this study proposes a machine-learning ensemble workflow to model diet and gut microbiome associations, and characterize patterns that are associated with depressive symptoms. The dataset included 88 participants, of which 15 with depressive symptoms. Microbial features were retained if consistently found by ANOVA, mutual information, and random forest importance. Dietary, microbial, and combined features were then evaluated using logistic regression (LR), support vector machine (SVM), and balanced random forest (BRF). We found that the SVM approach with combined features achieved the highest mean ROC-AUC (0.64), outperforming individual feature sets. Prioritized features included Lachnospiraceae-related taxa and cholesterol intake. Functional predictions suggested that Lachnospiraceae_ND3007_group and Eisenbergiella may carry genes involved in bile-acid transformation and acetate and butyrate metabolism, pathways that have been proposed to link the gut microbiome to mood regulation. Serum metabolic measurements demonstrated a moderate positive correlation between Rhizobiales_unclassified and acetate (Spearman ρ = 0.372, FDR q = 0.013), although no association was found between acetate and depression status. Overall, our findings show that ensemble machine learning can capture complex patterns linking diet, the gut microbiome, and depressive symptoms. Short-chain fatty acid (SCFA) measurements provided additional biological context, while no direct association with depressive symptoms or diet were observed.
Source: Preprints.org
Language: English
Topic: Gut microbiota and health
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7221636811