Hi,
I've banged my head long enough with this one, so I hope someone can help. I am
trying to write a routine to permute the structure of a tree. Note, this is
different than phyloshuffle() in phylotools, which just shuffles tips -- in
contrast, I am trying to permute the structure of the tree. In this case I
cannot resample the original data because these trees are manually constructed,
yet I want a way or doing some statistics on them. If there is a package doing
this that I've missed? I would like to be able to specify a number of
iterations (easy) and the routine would repeatedly switch random pairs of
clades within the tree (not so easy, apparently). If the routine is run for
many iterations, the resulting tree would be completely random. This seems
simple, but given the dependencies in the tree, you can't pick just any two
nodes; for example, the second node cannot be a descendant of the first node.
Also, it seems like I need to test that the two subtrees are structurally !
independent, but I'm not certain exactly what these criteria are. They don't
appear to have to be the same distance from the root in all cases. The next
problem is how to do this in ape. I'm reasonably familiar with the phylo()
object, but obviously since I'm posting this, I haven't been able to make it
work. Especially, the bind.tree() step.
Thanks much for any help,
Sean
## Tree Permutation
## Note, code does not work correctly!
library(ape)
library(distory)
atree<-read.tree(text="((L1496,L1499),((L1498,L1490),((L1497,L1494),((L1493,L1492),(L1503,(L1502,(L1212,L1501)))))),L1489);")
atree<-makeNodeLabel(atree, prefix = "")
atree<-compute.brlen(atree,1)
n=1 # No. itterations
plot.phylo(atree, show.node.label = TRUE)
ptree<-atree # a copy to permute
for (i in 1:n){
# Pick the first node from the tree
a.node<-as.numeric(sample(ptree$node,1))
a.node <- 2 # testing
# This creates a subtree to sample the second branch from
extract.a<-extract.clade(ptree,node=as.character(a.node))
drop.a.node<-drop.tip(ptree,extract.a$tip.label, trim.internal=F) #trim F to
leave a place to attach other subtree
plot(extract.a, show.node.label = TRUE)
plot(drop.a.node, show.node.label = TRUE)
# Pick a node from somewhere ealse in the tree
b.node<-as.numeric(sample(drop.a.node$node,1))
b.node <- 5 #testing
# This creates a subtree to sample the second branch from
extract.b<-extract.clade(drop.a.node,node=as.character(b.node))
drop.b.node<-drop.tip(drop.a.node,extract.b$tip.label, trim.internal=F)
plot(extract.b, show.node.label = TRUE)
plot(drop.b.node, show.node.label = TRUE)
# drop.b.node contains whatever is left after clipping out the two subtrees
# now I want to reassemble the tree switching extract.a and extract.b
bind.tree.1<-bind.tree(drop.b.node,extract.a,where=b.node) # doesnt work?
plot(bind.tree.1, show.node.label = TRUE)
bind.tree.2<-bind.tree(bind.tree.1,extract.b,where=a.node)
plot(bind.tree.2, show.node.label = TRUE)
#copy the results
ptree <- bind.tree.2
}
#plot the permuted tree
plot(ptree, show.node.label = TRUE)
#END
*************************************
Sean S. Downey, Ph.D.
Institute of Archaeology
University College London
31-34 Gordon Square
London
WC1H 0PY
s.dow...@ucl.ac.uk
http://www.homepages.ucl.ac.uk/~tcrnssd/
UK Cell: +44 (0)7542 41 0077
Belize Cell: +501 636-2939
_______________________________________________
R-sig-phylo mailing list
R-sig-phylo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
