I have five species trees run using starBeast and Beauti. Here is a link to the paper discussing both, http://journals.plos.org/ploscompbiol/article/file?type=supplementary&id=info:doi/10.1371/journal.pcbi.1003537.s006 Each of these trees has the same individuals, and each locus used (10 loci per tree) has individual models set to it. All the trees used a relaxed log clock normal. I would like to be able to combine these trees into one. I have attempted to find a way to do this within Densitree and starBeast, but to no avail. When all loci are run with all trees, the program simply dies, leading me to believe that running the trees together will all parameters is computationally impossible for my hardware. Is there a way to combine these trees?
1
-
$\begingroup$ I am not familiar with trees but what are these software that you use? Could you provide a link to see what are they capable? What is a "relaxed log clock normal"? Why don't you use all 30 loci to build a single tree by combining the individual models? What do you mean that is computationally impossible: how much time do you expect to take (with how many cpus?) ? $\endgroup$– llrsCommented Feb 12, 2018 at 21:14
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Although I've never tried it myself, I think what you need can be achieved through the LogCombiner program from the BEAST package -- assuming the species trees files are comparable (ran with same parameters, same species names, etc.) I am also assuming that you meant 'five species tree files', each file with the posterior sample of species trees.
However, I do believe you should try to run all loci within the same analysis, or at least do some prior check to see if the distributions between data sets are comparable. LogCombiner assumes independent runs from the same data.
I would also recommend taking a look (and following) the BEAST google group (which includes starBEAST discussions).
