A Bayesian Joint Model for Compositional Mediation Effect Selection in Microbiome Data

TL;DR

This paper introduces a Bayesian joint modeling approach for compositional microbiome data that enables simultaneous identification and quantification of direct and indirect mediation effects, addressing high-dimensional and overdispersed data challenges.

Contribution

It presents a novel Bayesian framework for compositional mediation analysis that jointly models effects and quantifies uncertainty, improving upon existing methods.

Findings

Outperforms existing methods in simulation studies for mediation effect selection.

Successfully applied to mouse microbiome data to identify antibiotic treatment effects.

Provides comprehensive uncertainty quantification for causal mediation effects.

Abstract

Analyzing multivariate count data generated by high-throughput sequencing technology in microbiome research studies is challenging due to the high-dimensional and compositional structure of the data and overdispersion. In practice, researchers are often interested in investigating how the microbiome may mediate the relation between an assigned treatment and an observed phenotypic response. Existing approaches designed for compositional mediation analysis are unable to simultaneously determine the presence of direct effects, relative indirect effects, and overall indirect effects, while quantifying their uncertainty. We propose a formulation of a Bayesian joint model for compositional data that allows for the identification, estimation, and uncertainty quantification of various causal estimands in high-dimensional mediation analysis. We conduct simulation studies and compare our method's mediation effects selection performance with existing methods. Finally, we apply our method to a benchmark data set investigating the sub-therapeutic antibiotic treatment effect on body weight in early-life mice.