FIGURE 1.

Intuition behind tascCODA. (A) A multifurcating tree structure $\mathcal{T}$ with internal nodes N1, N2, N3, and tips T1 …T6. tascCODA decides whether modeling the change of abundance of a subtree (e.g. nodes T5, T6 - gold). as a common effect at their common ancestor (e.g., N3 - red) is preferable. The blue nodes T1, N1, and N2 are reference nodes in this example. (B) Ancestor matrix of the tree in (A). (C) Example dataset where the abundances of T5 and T6 increase in the same way between conditions (relative to the reference T1). Here, a group-level effect on N3 would be the preferred option. (D) Plate representation of the tascCODA model. Grey squares indicate fixed parameters and input variables that are either part of or directly calculated from the data. The grey circle represents the output count matrix, white circles show latent variables.