B0801
Title: Novel regression models for microbiome data
Authors: Christian L. Mueller - Simons Foundation (United States) [presenting]
Abstract: Targeted amplicon sequencing data, including 16S rRNA and ITS sequence data, are inherently compositional in nature. Using these data for regression tasks thus requires non-standard regression models that take compositionality into account. In addition, typical microbiome data are overdispersed and zero-inflated. To alleviate the challenges associated with these data, we present novel regression models for microbiome data that jointly model the underlying regression and scale vectors under a wide range modeling assumptions. The corresponding model estimation task can be formulated as non-smooth convex optimization problem which can be solved efficiently using a novel proximal algorithm formulation. We show improved prediction performance compared to state-of-the-art methods for regression tasks arising in microbiome data analysis from host-associated and environmental amplicon data.
