Dear R users, I am trying to apply some form of phylogenetic comparative method (PIC, PGLS) to my morphological data so that I can run multivariate statistical tests like MANOVA, but I am stuck on this issue & could use some advice--
I have 11 morphological variables for 52 species, for which I also have a dated phylogeny. I ran traditional stats on the data to test if species means were significantly associated with habitat using MANOVA, and then did pairwise comparisons of species means between each habitat type using F tests. I found lots of morphological overlap. I also ran Pagel's lambda test on the species means and found significant phylogenetic signal. So what I would like to do is adjust for phylogeny (again, using PIC or PGLS... at this point I am just considering a Brownian motion model of evolution), and rerun the analysis using the phylo-adjusted means. I tried using the pic command in APE pic(x, phy, scaled = TRUE, var.contrasts = FALSE) but the problem is that I get out 1 less contrast out than terminal taxa, which makes sense since its based on internal nodes. I also know about this command picreconstruction <- ace(wingL, geotree, type="continuous", method="pic") but I think that would lead to the same problem. Is it possible to get a phylo adjusted score for each species to use in further statistical tests? I'm fairly new to this so any advice would be well appreciated. Many thanks, Christy Hipsley Museum für Naturkunde, Berlin, Germany _______________________________________________ R-sig-phylo mailing list [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
