Ecological Archives E096-256-A2
Anne Kempel, Mialy Razanajatovo, Claudia Stein, Sybille B. Unsicker, Harald Auge, Wolfgang W. Weisser, Markus Fischer, and Daniel Prati. 2015. Herbivore preference drives plant community composition. Ecology 96:2923–2934. http://dx.doi.org/10.1890/14-2125.1
Appendix B. Manual to compute phylogenetic independent contrasts for variables differing in the strength of phylogenetic signal, using the statistical software R.
We calculated phylogenetic independent contrasts (PICs) for all variables. However, as some variables are more or less phylogenetically constrained than others, we first created a specific tree for each of our measured variables, depending on the strength of the phylogenetic signal of a variable. To do so, we performed a phylogenetic regression (PGLS) for each variable to obtain a parameter for the strength of phylogenetic signal in the residual variation (Grafen’s rho):
m<- gls(variableA~1, corr=corGrafen (0.5,tree,fixed=FALSE))
We then used the obtained Grafen’s rho for each variable (parameter for the strength of the signal in the residual variation) to adjust the branch length of the phylogenetic tree of that variable:
new.tree.variableA<-compute.brlen(tree, power=rhoA)
Finally, we computed PICs for each variable, using the adjusted phylogenetic trees.
pic.variableA<-pic(variableA, new.tree.variableA)