I am working on my downstream analysis and could use some help clarifying the difference between the Unweighted vs Weighted unifrac test. As I understand, weighted (quantitative) accounts for abundance of observed organisms and unweighted (qualitative) is based on their presence or absence.
What is meant by abundance? Is it the number of reads assigned to an ASV?
Correct. Weighted UniFrac incorporates these abundances when calculating shared/unshared branch lengths to calculate distance, so the impact of low-abundance features is diminished.
In practice, this means that Weighted UniFrac is useful for examining differences in community structure; Unweighted UniFrac is more sensitive to differences in low-abundance features. So both are useful to interpret together and there are cases where both, only one, or neither may yield significant differences between experimental groups.
This article goes into more detail comparing when/where/how to use these metrics.
Nothing to worry about — sampling depth (read count) is normalized by even sampling prior to calculating beta diversity in the core-metrics actions, and other normalization strategies will be offered in the future.
This may remind you of another “no no” in the amplicon NGS world: the inadmissability of rarefying data prior to some analyses. But there is nothing to worry about here, either. UniFrac is not really negatively impacted by rarefying, and Unweighted Unifrac requires rarefying. See more discussion of this topic here.