B0835
Title: Tree-based models for high-dimensional compositional data in microbiome studies
Authors: Zhuoqun Wang - Duke University (United States) [presenting]
Abstract: The human gut microbiome is associated with various diseases and health outcomes. A key characteristic of microbiome compositional data is its large and complex cross-sample heterogeneity. Appropriately accounting for these variance components is critical for several common inference tasks, including identifying latent structures, carrying out hypothesis testing on cross-group differences, and modeling dynamics, but is complicated by the key features of microbiome compositional data, including high-dimensionality, sparsity, and compositionality. These characteristics incur the need for structural constraints on covariance modeling while maintaining the analytical and computational tractability of the resulting models and methods. We present recently proposed methods that aim to utilize a tree structure -- namely the phylogeny of the microbial species -- to incorporate flexible covariance components while maintaining computational scalability. In particular, we present probabilistic models for microbiome compositional data based on the logistic-tree normal (LTN) distribution and demonstrate their wide applicability in a range of applications, including mixed-effects modeling, covariance estimation, and differential abundance analysis.
