Hi Roland.
Thanks for your interest in this function and method. The method is
described in an article that is now available as an "Accepted Article"
online (http://dx.doi.org/10.1111/j.1558-5646.2011.01435.x).
Unfortunately, this was posted online by 'Evolution' without its
supplementary appendix (which contains details on running and
interpreting the function evol.rate.mcmc()), but I would be happy to
send this to you off-list.
evol.rate.mcmc() implements a method in which MCMC is used to identify
the phylogenetic position of a shift in the evolutionary rate for a
continuously valued character. It can also be used to obtain estimates
of the evolutionary rates tipward and rootward of this shift.
What you are looking at is the "raw" posterior sample from this
analysis. Depending on how long you run the MCMC you should have 100s
or 1000s of rows in $mcmc & 100s or 1000s of elements in $tips. To make
sense of this, you should use two different functions: minSplit() and
posterior.evolrate().
minSplit() takes the tree and $mcmc as input. You should first decide
how many generations (and thus how many rows of $mcmc, which will depend
on sampling frequency) you want to exclude as burn-in. It returns the
shift-point that minimizes the summed or the sum of squared distances to
all the other sampled points in the posterior.
posterior.evolrate() pre-processes the posterior sample so that you can
make sense of it. It takes as input the tree, the average shift point
(say, the output from minSplit()), $mcmc, and $tips. It returns a
matrix containing a pre-processed posterior sample of evolutionary rates
tipward and rootward of the median shift point which can then be
analyzed in a typical way. (For instance, you can compute parameter
estimates as the mean or median; as well as HPD intervals and ESSs
using, for instance, the MCMC diagnostics package "coda.")
I hope this is helpful. Thanks again for giving the method a try.
Please don't hesitate to contact me for further clarification.
Best of luck.
- Liam
--
Liam J. Revell
University of Massachusetts Boston
web: http://faculty.umb.edu/liam.revell/
email: liam.rev...@umb.edu
blog: http://phytools.blogspot.com
On 8/12/2011 11:07 AM, Roland Sookias wrote:
Hi
Can anyone help me interpret the output from phytools' rate.evol.mcmc
function? I get, as it says I should, "results from the MCMC run" and
a "tips list of stips in state sig(1)^2 for each sampled generation of
MCMC (to polarize the rate shift).", but I'm not sure what I'm looking
for practically to see if/when there have been rate shifts.
The mcmc list is like this:
2 200 0.004979113 0.002156864 1.889112 131 3.6386714 -38.68685
3 300 0.008243820 0.006182640 1.964227 97 19.7978315 -36.52512
And the tips like this:
$tips[[5]]
[1] "Effigia_okeeffeae"
Cheers
Roland
_______________________________________________
R-sig-phylo mailing list
R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
_______________________________________________
R-sig-phylo mailing list
R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
