Hi Dan,
Even though this is well-trodden ground in some respects (by the likes of
TG and others), I have often-enough fielded questions about
size-correction and principal components analysis on species data (or
encountered those who see nothing wrong with ignoring phylogeny in said
procedures) that I recently wrote and submitted a Brief Communication to
'Evolution' on the topic entitled "Statistical transformations and trees:
Phylogenetic size-correction and principal components." In the article, I
provide the mathematical details of these procedures, as well as simple R
and Matlab code in an appendix. I also analyzed the variance and type I
error associated with ignoring phylogeny in these preliminary corrections
or data reduction procedures. Although the effect is not incredibly
severe, it is large enough to be of significant concern. I'm happy to
send this manuscript to anyone who is interested. As I said, it is
presently in review, but I'm encouraged by this discussion thread that
there seems to be considerable interest in this area.
I also performed my type I error analyses using stochastic pure-birth
trees. Obviously, the effect on type I error will be increased for more
realistic birth-death trees (in which there are more shorter branches
towards to tips of the tree).
Thanks! - Liam
Liam J. Revell
Department of Organismic and Evolutionary Biology
Harvard University
web: http://anolis.oeb.harvard.edu/~liam/
email: lrev...@fas.harvard.edu
On Mon, 6 Apr 2009, Luke Harmon wrote:
You should reply to the list.
Sent from my iPod
Begin forwarded message:
From: "Brian O'Meara" <bcome...@nescent.org>
Date: April 6, 2009 9:04:00 AM PDT (CA)
To: Dan Rabosky <dl...@cornell.edu>
Cc: r-sig-phylo@r-project.org
Subject: Re: [R-sig-phylo] size-free morphometrics in phylogenetic
framework
Hi, Dan et al. I've been worried about this for PCA. Mesquite has an
"evolutionary PCA" approach, but as far as I know, this hasn't been
published anywhere. Assuming the method implemented in Mesquite is done
properly (which I think rather likely), it'd be relatively simple to
compare the effect of the evolutionary PCA vs conventional approach on real
and simulated datasets.
Brian
On Apr 6, 2009, at 11:53 AM, Dan Rabosky wrote:
Howdy folks-
I think I'm opening this up for discussion, rather than pointing to a
specific problem.
Suppose I'm conducting an analysis of phenotypic evolution and need
size-free species measurements. A standard approach might be to take
species mean values, then regress those values on some index of size
(e.g., snout-vent length), and treat the residuals as size-free
measurements for further analysis. However, I've also seen explicit
incorporation of phylogenetic info into the estimation of size-free
variables. For example, Collar et al (Evolution, online early, DOI:
10.1111/j.1558-5646.2009.00626.x) used PIC to estimate regression
slopes of traits on size, then forced these slopes on regressions of
raw species trait values and used the residuals as size-free variables.
Do we have strong feelings about why/whether the latter is the
approach that should be used? My intuition, probably incorrect, is
that the slopes should be relatively robust regardless of whether
phylogeny is used, but not the degrees of freedom.
And if this is a serious problem, wouldn't PCA and geometric
morphometric approaches also be affected by the assumption that
species values represent independent data points?
~Dan
Cornell University
http://www.eeb.cornell.edu/Rabosky/dan/main.html
[[alternative HTML version deleted]]
_______________________________________________
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
_______________________________________________
R-sig-phylo mailing list
R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
