Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 385 - Recent Advances in Bayesian Methods for the Analysis of Microbiome Count Data
Type: Invited
Date/Time: Wednesday, August 5, 2020 : 1:00 PM to 2:50 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #309207
Title: Efficient Bayesian Estimation of Microbiome Association Networks
Author(s): Christine B. Peterson* and Nathan Osborne and Marina Vannucci
Companies: The University of Texas MD Anderson Cancer Center and Rice University and Rice University
Keywords: Microbiome; Graphical model; Network inference; Bayesian modeling; Variational inference

We propose a novel method to simultaneously estimate network interactions and associations to relevant covariates for microbiome data. Learning these relationships is challenging due to the unique structure of microbiome data, where the abundances of operational taxonomic units (OTUs) are counts subject to a sum constraint. We have developed a Bayesian hierarchical model for the estimation of these association networks which not only handles the compositional nature of the data, but also allows us to select relevant covariates in a joint framework. To obtain posterior estimates, we rely on an efficient algorithm based on expectation conditional maximization. We illustrate the proposed method through both simulation studies and an application to real data from the MOMS-PI study on the abundance of vaginal microbes and cytokines, providing insight into mechanisms of host-microbial interaction during pregnancy.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program