How to express what the ingroup is?

[jar398]

Many trees have an ingroup/outgroup distinction, which affects both data collection and biological conclusions. Knowing the ingroup, when one is known, is essential for meaningful scientific use of trees. (This issue has turned out to be central in the Open Tree of Life synthesis application.)

[hlapp]

@jar398 if Open Tree has decided on a tentative way to support this, can you add that here, and feel free to add your thoughts about that, such as how well it has worked for you.

[jar398]

Not sure of the details but we're putting a boolean property 'is an ingroup' on the nexml node as metadata, and we're making sure that when this is done, the ingroup is that node and anything reachable by going forward through directed edges in the representation. That is, not only do we label the ingroup, but we assume (and enforce) that the representation rooting of the tree is 'above' the ingroup node. (you have to rule out cases like ((a,b)z,(c,d)ingroup)root where the tree is assumed 'unrooted' meaning that the ingroup is ((a,b)z)ingroup and (c,d) is the outgroup - it's ambiguous because the tree is unrooted.)