polytomies in the trees

[Grygoriy Zolotarov]

I have realized that many of the gene family trees generated by GeneRax have polytomies (i.e. nodes with length=0.000001). 

Is there any way to get rid of these? 

Cheers,

Grisha

[Benoit Morel]

Hi Grisha,

I am so sorry, I thought I had replied to this thread!

Nodes with length 0.000001 can be seen as polytomies with tools such as raxml or iqtree, but not necessarily with GeneRax. The branch length is the expected average number of substitutions per site. Almost null branch lengths mean that there were no substitutions along this branch.

For tools that only take as input the sequences, this cannot be resolved (no substitution = no information) and this is indeed a polytomy. But for tools that also use the species tree, those nodes can be resolved by using the species tree topology (or at least, there are good reasons to think that this resolution is more plausible than the alternative resolutions). The branch lengths are still ~0, because they still represent the expected average number of substitutions (which is indeed 0). It doesn't mean that the three children diverged simultaneously in units of time.

Cheers,

Benoit