Incorporating divergence time uncertainty from MCMCtree

[Skip]

Hi,

I using MCMCtree to estimate divergence times for a phylogenomic dataset with 190 species and I'm currently running a number of comparative phylogenetic methods on the MCC tree output. I was wondering if anyone knew of a way to map the 95% HPD estimates from the posterior onto the tree in newick format so that these can be read into R as a multiphylo object. This would allow me to test the robustness of models that are dependent on the timing of divergence (i.e.- TESS and DDD). Hopefully, this makes sense. Basically, I'm looking for some way of coercing the MCMCtree output to produce something analogous to the '.trees' output from BEAST. If this is not an option or this question comes off as utter nonsense, please let me know.

Best,

Skip

[Ziheng]

The program should produce a file with a name like FigTree,tre, which is in the newick format and readable by Figtree.

it's probably the same format as produced by beast.  the end of the file may have some instructions about options in Figtree, to create node bars, time axis etc.

ziheng

[Pierre Arnal]

Hi Ziheng,

If think Skip was more thinking about a way to sample X trees from the posterior distribution. I am actually looking for a way to do so. If you have any advice it would be amazing !

[Ziheng]

to generate a posterior sample of the trees (or the same tree with different divergence times more precisely) from the mcmctree sample, you will have to do some coding.

suppose you can read in the tree from the mcmctree control file.  then scan the mcmc.txt file (the mcmc sample file) to read in the node ages, and assign the node ages to the nodes in the tree.  perhaps with the ape package in R this should not be hard?

ziheng