Boyu Ren (Harvard T.H. Chan School of Public Health)

A Bayesian Nonparametric model for microbiome data analysis

We develop a statistical model to analyse microbiome profiling data based on sequencing of genetic fingerprints in 16S ribosomal RNA. The analysis allows us to quantify the uncertainty in  ecological ordination and clustering methods commonly applied in microbiome research. In addition, it can be extended into a framework for association studies when sample characteristics are available. The method is based on the estimation of the underlying microbial distribution in experimental samples using a dependent Dirichet Process prior in which dependence is expressed through the combination of low-dimensional latent features and observed sample covariates. This type of model is advantageous for several reasons. First, information is borrowed across samples to estimate underlying microbial distributions. Second, the nonparametric nature of the model avoids the artefacts of truncation and rarefaction techniques. Lastly, the Bayesian framework mitigates the effects of multiple testing for associations between covariates and species abundance and other hypotheses of interest.