If you want to do a Bayesian reconstruction for continuous traits with a
strong prior on the root - as Brian suggests - you can do this in the
phytools function anc.Bayes. The prior means & variances need to be
provided in anc.Bayes(...,control=list(pr.var,pr.mean)). Unfortunately,
you need to put priors on everything (or nothing, in which a pretty
uninformative prior is used). If you decide this is what you want to do
& have difficulty figuring it out, please let me know.
All the best, Liam
Liam J. Revell, Assistant Professor of Biology
University of Massachusetts Boston
web: http://faculty.umb.edu/liam.revell/
email: liam.rev...@umb.edu
blog: http://blog.phytools.org
On 4/19/2013 4:56 PM, Brian O'Meara wrote:
We have a method in R, TreEvo, that can allow you to do things like have a
model where the optimum shifts through time according to some factor (like
external data on climate) or where there are constraints that change
through time: it's up on R-forge, and you can install the package, but it
hasn't been published yet and so could have problems that won't appear
until peers have looked at it more thoroughly. If you're fine waiting to
publish with it until it's been vetted and is in press, feel free to play
with it.
Of course, another thing you could do is throw a prior on the root state
based on your knowledge of the climate then and then do a Bayesian
reconstruction. I don't know of any packages that do this at the moment,
but it wouldn't be hard to write.
Best,
Brian
_______________________________________
Brian O'Meara
Assistant Professor
Dept. of Ecology & Evolutionary Biology
U. of Tennessee, Knoxville
http://www.brianomeara.info
Students wanted: Applications due Dec. 15, annually
Postdoc collaborators wanted: Check NIMBioS' website
Calendar: http://www.brianomeara.info/calendars/omeara
On Fri, Apr 19, 2013 at 4:00 PM, Eliot Miller <eliotmil...@umsl.edu> wrote:
Hello all,
I've read some past posts and tried to figure this one out myself. It
appears that the short answer is if there are not some extinct lineages in
the tree with known trait values, it's difficult to do.
I have a data set of the current climatic niches of a continental radiation
of taxa. If I reconstruct the ancestral state of the root using ace() with
REML, I get a semi-believable answer, though obviously with wide CIs. My
gut instinct is that the absolute value that is returned is slightly less
than it should be.
The issue is that I know the climate of the continent has changed radically
(but at least consistently in one direction!) during the course of the
radiation of the clade. An educated guess would put the magnitude of the
change at 1200 to 550 mm rain per year over 60 million years. I'd guess
that most lineages have felt this aridification as a slow and steady
"selective" pressure to "figure it out" in increasingly dry conditions
(there's probably also been a lot of extinction but let's skip that point
for now).
So, while I don't have extinct lineages with known trait values, I do have
what I'd be comfortable modeling as a continuous selective pressure
throughout the course of the radiation. I'm interested to see how doing
this would change the actual absolute value returned by the ancestral state
reconstruction. Any tips on how this might be best effected?
Final note, my tree is from uncorrected molecular distances, so it's not
ultrametric. The analysis I'm trying to run here is tangential to my main
focus, so I'm ok with "making" the tree ultrametric with something like
chronopl() just out of curiosity to see what comes out. Someday I'll have a
time calibrated tree and then I can re-run this analysis.
Best wishes,
Eliot
[[alternative HTML version deleted]]
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at
http://www.mail-archive.com/r-sig-phylo@r-project.org/
[[alternative HTML version deleted]]
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
_______________________________________________
R-sig-phylo mailing list - R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/r-sig-phylo@r-project.org/
