Penalized Regression Methods for Microbiome Compositional Data by Prof. Anna Plantinga, Statistics Colloquium, November 14, 1:10 - 1:50 pm

Penalized Regression Methods for Microbiome Compositional Data by Prof. Anna Plantinga, November 14, 1:10 – 1:50 pm, Statistics Colloquium

Abstract:  The human microbiome plays a vital role in maintaining health, and imbalances in the microbiome are associated with a wide variety of diseases. However, analysis of microbiome data is challenging because the data are high-dimensional and compositional with underlying biological structure. In this talk, we present two penalized regression methods for estimation and prediction with high-dimensional compositional data. Because phylogenetic similarity between bacteria often corresponds to shared functions, our first model incorporates phylogenetic structure into a penalized regression model for constrained data. We then propose a model that exploits this phylogenetic structure to use partial information when feature sets differ between the model-building and prediction datasets. We evaluate the performance of these methods through extensive simulation studies and apply them to studies investigating the association of body mass index with the gut microbiome.