omicverse-microbiome-phylogeny

To build a phylogenetic tree for UniFrac analysis, you construct an ASV-level tree using MAFFT alignment and FastTree under a GTR plus gamma model. This process creates the tree structure required to compute phylogenetically-aware beta diversity metrics.

Attaching a phylogenetic tree to AnnData involves validating tree tip names against ASV features and pruning mismatches. The validated tree is stored in adata.uns and adata.obsp, enabling direct computation of cohort diversity metrics.

You can compute Faith PD for alpha diversity, alongside unweighted UniFrac and weighted UniFrac for beta diversity, from 16S amplicon data. These metrics provide phylogenetically-aware insights into microbiome cohort differences.

Yes, MAFFT and FastTree are required to construct the ASV phylogenetic tree needed to calculate Faith PD and UniFrac metrics. Without this tree-building step, you cannot compute phylogenetically-aware diversity from 16S sequences.

The best way to compare UniFrac against Bray-Curtis is to compute both unweighted and weighted UniFrac from a tree-augmented AnnData object. This allows you to contrast phylogenetically-aware beta diversity with standard abundance-based metrics.

Tree tip validation fails when attaching phylogenetic trees to AnnData due to mismatches between tree tip names and ASV feature names. The Skill addresses this by validating tip names and pruning mismatches when appropriate before attachment.