[R-sig-phylo] GEE warning message: Working correlation estimate not positive definite

2011-04-01 Thread jonathan ROLLAND
Hello, I'm new in gee, and I want to detect AFLP loci (y) (column of 0 and 1) under the selection of one environmental variable (x). I'm using gee(locus~x,id=data$sites ,corstr=exchangeable, family=binomial). package (gee) I'm using gee because of the autocorrelation in my experiment design

Re: [R-sig-phylo] How to detect phylogenetic signal (lambda) in one unscaled trait?

2011-04-01 Thread Marguerite Butler
Hi Folks, After that very long-winded answer I still managed to leave out something. I want to reinforce what Ted and Joe said regarding absolute vs. log scales. Most comparative methods machinery as currently implemented work with absolute scales (the random component of the evolutionary

Re: [R-sig-phylo] Adding a taxon to a pre-exisiting tree

2011-04-01 Thread Liam J. Revell
Hi Matthew. I don't doubt that other members of the list have better suggestions, but it is possible to add a tip in all possible places using bind.tree() in ape. For instance, starting with a random unrooted tree with, say, 4 taxa: tree-rtree(n=4,rooted=FALSE,br=rep(1,5)) # create a 5th

Re: [R-sig-phylo] Adding a taxon to a pre-exisiting tree

2011-04-01 Thread Liam J. Revell
Of course, we could generalize my preceding suggestion with the following function: add.everywhere-function(tree,tip.name){ tree-unroot(tree) tree$edge.length-rep(1,nrow(tree$edge)) new.tip-list(edge=matrix(c(2,1),1,2),tip.label=tip.name, edge.length=1,Nnode=1)