search

richelbilderbeek / pirouette_example_16

pirouette example 16
https://github.com/richelbilderbeek/pirouette_examples
0 stars 0 forks source link

include option to analyze trees with extinct branches #2

Open thijsjanzen opened 4 years ago

thijsjanzen commented 4 years ago

Node substitutions also arise along branching points that lead to extinct species, these so called 'hidden nodes' (because they don't show up in the reconstructed tree), are an important aspect of the node substitution model.

Currently, pirouette does not support alignment generation using a tree including extinct branches (the nodesub function do - they generate the full alignment, and then purge the extinct sequences). I'm not entirely sure where in the pirouette machinery to correct for this.

richelbilderbeek commented 4 years ago

I'm not entirely sure where in the pirouette machinery to correct for this.

AFAICS, in short, one cannot.

I do know babette has experimental support for tip dating, allowing to use a DNA sequence at the moment a species goes extinct. Would that work (it is yet untested) and be desired, pirouette should be able to more-or-less do so...

thijsjanzen commented 4 years ago

Well, the extinct branches are only to be taken into account for generating the twin alignment - not for generating the twin tree (although that would be interesting as well, but probably enough to base an entire PhD on)

richelbilderbeek commented 4 years ago

Hmmm, now it starts sounding perfectly doable!

I'd say (without to much more thought) it can be done, by simply relaxing some assumptions of pirouette. Just removing some ifs should allow for what you want.