phyloMDA

phyloMDA performs phylogeny-aware statistical analysis of microbiome sequencing data to model multivariate compositional abundances and infer associations using accompanying phylogenetic trees from modern sequencing technologies.

Key Features:

• Phylogeny-Aware Analysis: Integrates phylogenetic trees accompanying microbiome datasets to account for shared evolutionary history when analyzing community patterns across sample sites, conditions, and treatments.
• Dirichlet-Tree Multinomial Model: Implements a Dirichlet-tree multinomial model for multivariate abundance data that handles the compositional nature of microbiome datasets.
• Empirical Bayes Estimation: Applies tree-guided empirical Bayes estimation to improve inference of microbial compositions using priors derived from phylogenetic structure.
• Tree-Based Multiscale Regression Methods: Provides tree-based multiscale regression methods using relative abundances as predictors to model relationships across phylogenetic scales and resolutions.

Scientific Applications:

• Host-Associated Community Dynamics: Characterizing dynamics and composition of host-associated microbial communities using compositional abundance modeling.
• Evolutionary Impact Studies: Investigating how evolutionary history influences microbial interactions and functions by incorporating phylogenetic structure.
• Treatment and Condition Effects: Assessing effects of conditions or treatments on microbiome composition while accounting for phylogeny.

Methodology:

Dirichlet-tree multinomial modeling, tree-guided empirical Bayes estimation, and tree-based multiscale regression applied to relative abundance (compositional) microbiome data with phylogenetic trees.