I am looking for main differences between two R Objects constructed for analysis of microbiome data: phyloseq object and TreeSummarizedExperiment object

First, I will provide a little information about both objects and then ask the questions later, in order to avoid repetitions of familiar points in the discussion.

1. Phyloseq:

   • Purpose: Phyloseq is a popular R package designed for construction of a phyloseq object which can used for the analysis and exploration of microbiome sequencing data.
   • Functionality:
     • It provides tools for importing, manipulating, and visualizing microbiome data.
     • Phyloseq handles data in a tabular format, where rows represent samples (e.g., microbial communities from different individuals or environments), and columns represent features (e.g., bacterial taxa or operational taxonomic units).
     • Users can perform diversity analyses, community comparisons, and visualization of microbial abundance.
   • Data Structure:
     • Phyloseq primarily uses a data frame-like structure to store sample metadata, feature abundance data, and taxonomic information.
   • Typical Workflow:
     1. Import raw sequencing data, Construct phylogenetic trees (if required).
     2. Process and filter the data (e.g., rarefaction, normalization).
     3. Compute alpha and beta diversity metrics.
     4. Visualize results (e.g., bar plots, heatmaps).

2. TreeSummarizedExperiment:

   • Purpose: TreeSummarizedExperiment is an extension of the SingleCellExperiment class, specifically tailored for microbiome data with hierarchical structures.
   • Functionality:
     • It combines rectangular experimental data (similar to SingleCellExperiment) with hierarchical tree structures.
     • TreeSummarizedExperiment can store both the experimental results (assays) and the hierarchical relationships (trees) simultaneously.
     • It is particularly useful for microbiome data where samples are related through phylogenetic trees.
   • Data Structure:
     • TreeSummarizedExperiment includes additional slots beyond those in SingleCellExperiment:
       • rowTree: Represents the hierarchical structure on the rows (e.g., taxa).
       • rowLinks: Provides link information between rows of assays and the row tree.
       • colTree: Represents the hierarchical structure on the columns (e.g., samples).
       • colLinks: Provides link information between columns of assays and the column tree.
       • referenceSeq (optional): Stores reference sequence data per feature (row).
     • The tree slots (rowTree and colTree) require the tree to be an object of the phylo class.
   • Typical Workflow:
     1. Import microbiome data.
     2. Construct phylogenetic trees (if available).
     3. Create a TreeSummarizedExperiment object by combining data, trees, and links.
     4. Perform analyses that leverage both the rectangular data and tree structures.

I have mentioned the obvious differences above, because I am looking for advice from people with experience with these packages.

My goal is to perform Longitudinal analysis of microbiome data with regards to different parameters of alpha and beta diversity.

I am looking for differences between the two packages in terms of Statistical tests, methodological considerations, community support and programming difficulty.

I would also like to know what these two packages differ in terms of handling different models like Generalized Linear Mixed Models (GLMMs) or Zero-Inflated Binomial Models to account for overdispersion and handles normalization and p-value adjustment