Closed cerebis closed 7 years ago
Simply put, a community is currently defined as a set of (tree, profile) tuples -- which then generate a set of clades. There can be 1 or more clades in a community and clades are eventually merged when performing read simulation.
So defined, clade profiles are independent within a community. For those with low interactions, this might be sensible, but for those interacting (competing) a single profile may be more realistic.
Therefore, we need to add support for a single profile definition for a community.
Todo:
I will be removing per-clade profiles for the time being. This is implemented on branch:experimental.
05a8e4f7395be85769ae3c577b8aeb47012f2e01
Rather than a single distribution governing the profiles of all taxa with a given tree, it would seem that certain topological characteristics should dictate dependence. For instance, deeply branched sub-clade or recent radiations; there will be scenarios where constraint should be imposed on taxon abundance assignment (currently over the entire tree) within that sub-clade.