Related Experiment Video
Updated: Oct 15, 2025
09:49
Divergence of Root Microbiota in Different Habitats based on Weighted Correlation Networks
Published on: September 25, 2021
4.5K
A Bayesian Negative Binomial Hierarchical Model for Identifying Diet-Gut Microbiome Associations.
Alma Revers1, Xiang Zhang2, Aeilko H Zwinderman1
1Department of Epidemiology and Data Science, Amsterdam University Medical Center, Amsterdam, Netherlands.
Frontiers in Microbiology
|October 25, 2021
Summary
Incorporating phylogenetic relationships improves diet-microbe association analysis in gut microbiome studies. This Bayesian model enhances accuracy, especially with over-dispersed data, offering robust insights into diet
Area of Science:
- Microbiome Research
- Statistical Modeling
- Nutritional Science
Background:
- Human gut microbiota composition is crucial for health.
- Long-term diet interventions significantly influence gut microbiome structure.
- Understanding diet-microbe associations is key to personalized nutrition and health.
Purpose of the Study:
- To develop a statistical model integrating phylogenetic relationships of operational taxonomic units (OTUs) for diet-microbe association analysis.
- To evaluate the performance of a Bayesian hierarchical negative binomial (NB) model with phylogenetic information compared to existing methods.
- To assess the impact of phylogenetic relationships on the accuracy and variance of diet-microbe association estimates.
Main Methods:
- Utilized a Bayesian hierarchical negative binomial (NB) model incorporating phylogenetic relationships between OTUs.
- Regularized the dispersion parameter of the NB distribution by assuming a mean-dispersion association.
- Validated the model using simulation studies and applied it to data from the Healthy Life in an Urban Setting (HELIUS) study.
Main Results:
- The proposed model demonstrated improved performance (lower Mean Squared Error) in simulations with over-dispersed microbiome data compared to standard NB models.
- In the HELIUS study, incorporating phylogenetic relationships reduced posterior variances for some phylogenetic families.
- However, bias was introduced when OTUs within the same family were not similarly affected by dietary factors, leading to increased variances.
Conclusions:
- The Bayesian hierarchical NB model, accounting for mean-dispersion dependency and phylogenetic relationships, offers a robust approach for analyzing diet-microbe associations.
- Phylogenetic information can enhance the precision of diet-microbe association estimates, but its application requires careful consideration of the similarity of microbial responses within phylogenetic groups.
- This methodology provides a more nuanced understanding of how diet shapes the gut microbiome, with implications for public health and nutritional interventions.
