search

roblanf / sarscov2phylo

Global phylogenies of SARS-CoV-2 sequences
GNU General Public License v3.0
86 stars 22 forks source link

keep polytomies #10

Closed roblanf closed 4 years ago

roblanf commented 4 years ago

Right now I randomly-resolve polytomies in trees, simply to allow me to add TBE and FBP values.

However, it's preferable to keep polytomies for lots of reasons. So, there are two options:

  1. Collapse polytomies in the final trees
  2. Estimate TBE and FBP values in gotree (https://github.com/evolbioinfo/gotree) instead of in raxml.

I prefer the second option, because it's more parsimonious. All that's needed is to first test whether gotree can handle trees this large in a reasonable timeframe.

roblanf commented 4 years ago

Update: gotree works fine for estimating the supports. Takes ~10 minutes on ~50 CPUs, and uses ~25GB memory. It will depend a little how the memory scales as the trees increase in size, but we have plenty of room if we're starting out here.

e.g. commandline:

gotree compute support tbe -i global.famulti.fasttree -b global.fa_ft_replicates_multi.tree -t 55 -o global.fa_ft_tbe.tree
roblanf commented 4 years ago

done