|
Abstract:
|
Longitudinal microbiome studies can help delineate true biological signals from the high interindividual variability that is common in microbiome data. However, there are limited methods available for the dimension reduction of time course microbial count observations. Existing methods do not fully observe the distribution characteristics of such data types, namely, zero-inflation, compositionality, and overdispersion. We present a semiparametric quasi-likelihood model for the decomposition of longitudinal microbiome datasets, by generalizing existing approaches in tensor decomposition of Gaussian data. Optimization is performed through projected gradient descent additionally allowing interpretability constraints. We show through simulation studies our method is able to recover low rank structures from microbiome time course data, better than existing options. Lastly, we applied our method to two existing longitudinal microbiome studies, to detect global microbial changes associated with dietary/pharmaceutical effects and infant delivery modes.
|
