Bayesian graphical compositional regression for microbiome data
share
An important task in microbiome studies is to test the existence of and give characterization to differences in the microbiome composition across groups of samples. Important challenges of this problem include the large within-group heterogeneities among samples and the existence of potential confounding variables that, when ignored, increase the chance of false discoveries and reduce the power for identifying true signals. We propose a Bayesian graphical regression strategy to overcome these issues based on a Dirichlet tree multinomial (DTM) model for compositional data, which performs a tree-based decomposition of the original problem into a series of local testing problems. Each local problem is solved with a Bayesian regression model that adjusts for covariates and group structures. A graphical structure is used for linking the local models to achieve efficient information sharing across OTUs and taxonomic levels. We conduct an extensive simulation study to investigate the performance of the approach. The results show that the method outperforms several existing approaches in a number of representative settings. We then apply our method to the American Gut data to analyze the association of dietary habits and human's gut microbiome composition in the presence of covariates, and illustrate the importance of incorporating covariates in microbiome cross-group comparison.
READ FULL TEXT