Hi Miriam.
It is the correlation structure (phylogenetic signal) of the residual
error of y|x that matters, not the structure of the individual variables.
You can use an OU model for the correlation structure of the residual
error in PGLS. This would look something like this (if 'yourdata' is a
data frame containing variables x y):
lmOU-gls(y~x,data=yourdata,correlation=corMartins(1,tree))
residuals(lmOU) # to get the residuals
Finally, Brownian motion is a special case of OU when alpha=0. If you
have lots of data, OU may fit better than BM even though alpha is very
close to zero. So any judgement on how good or bad it will be to
assume BM depends on the fitted parameter values for alpha.
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 8/21/2013 12:52 PM, Miriam Santos-Herrera wrote:
Dear all,
Im an newbie in this world of comparative methods and excuse me in advance
if I am asking something very obvious.
I have performed phylogenetic size correction and a phylogenetic pca to my
data using phyl.resid and phyl.pca. My concern is that these functions
assume that my traits evolve following a constant rate BM across the tree.
Then when I fit the different models of evolution to my transformed
variables, I see that the best model is not a constant rate BM but a OU1
model.
If I understand things well, a BM model is a very bad model when there is
consistent selection towards a single optimum trait value.
My question is, which might be the effect of having assumed a wrong model
in these previous transformations of my variables?
and this lead me to the next question:
between not correcting your residuals and pca by phylogeny, and correcting
them despite of using a possibly wrong model, which option is the best? (or
the less worse).
any light on this would be very appreciated!
Miriam.
[[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/