Hi Oscar.
I'm not sure what you're trying to accomplish. By fitting
gls(y~Tree,correlation=Tree) you have the phylogeny in both the mean
structure & the error structure of your fitted model. This does not seem
like a good idea.
If you just want to measure the degree to which your data are
phylogenetically structured (i.e., the "effect" of phylogeny on your
trait - although this is probably not a good way of describing it
because your phylogeny cannot affect the trait, although history may
influence how variation is structured among your tip species), you
should probably just measure phylogenetic signal using one or multiple
methods. For instance, you could use my function phylosig in the
phytools package.
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 9/6/2013 1:22 PM, Oscar Valverde wrote:
Hi everyone,
I have been trying to analyze the effect of phylogeny on trait integration
among plant tissues. Initially I tried a non-phylogenetic approach using
species nested within clades as error term
model1<-lmer(log(Trait)~Tissue*Clade*Site+(1|Clade/Species),data=data)
then, I tried to remove the phylogenetic value using the residuals after
running a model with the samples fixed by clade in a pgls model and making
a new dataset
Tissueresid1<-residuals(gls(log(Tissue)~Clade,correlation=corPagel(1,tree),data=datatissue1))
Tissueresid2<-residuals(gls(log(Tissue)~Clade,correlation=corPagel(1,tree),data=datatissue2))
Tissueresid3<-residuals(gls(log(Tissue)~Clade,correlation=corPagel(1,tree),data=datatissue3))
dataphylo<-cbind(Tissueresid1,Tissueresid2,Tissueresid3)
then I run a new model and tried to compare the AIC for the same model
with the corrected matrix
modelphylo1<-lmer(log(Trait)~Tissue*Clade*Site+(1|Clade/Species),data=dataphylo)
The AIC of both models however are identical and but none of the factors
are significant anymore!. What I'm doing is wrong? Any suggestions of an
alternative approach to test the integration of tissues including the
phylogenetic relatedness among species?
I'll appreciate any comments, best regards
_______________________________________________
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/
