search

lmaurits / BEASTling

A linguistics-focussed command line tool for generating BEAST XML files.
BSD 2-Clause "Simplified" License
20 stars 6 forks source link

Take starting tree from Glottolog #101

Open Anaphory opened 8 years ago

Anaphory commented 8 years ago

I thought we had that capacity already and was looking for a reminder how we did it, but apparently we have no way to specify the subtree(s)* of the Glottolog tree spanned by the languages in the analysis as starting tree.

*) And I think dealing with a forest may be part of the problem

lmaurits commented 7 years ago

I suppose you are wanting to do this for an analysis where you are not imposing Glottolog monophyly constraints?

Anaphory commented 7 years ago

I can't remember – It could also have been (sounds more likely to me now) “use that tree only, don't change topology”, for which monophylies would be overkill.

lmaurits commented 7 years ago

I only asked because if you turn on monophyly constraints, then BEAST will make sure that the starting tree is consistent with them, so in effect you get a Glottolog tree with polytomies resolved at random - and since the starting tree has to be fully binary then this is the only thing I can imagine you were asking for. So you kind of get this for free unless you are not using monophyly constraints.

By using monophyly constraints and setting sample_topology=False you basically end up just sampling branch lengths for some randomly-resolved Glottolog tree.