Hi Kevin,
here are some bits of code, which should let you compute those numbers:
library(phangorn)
set.seed(3)
nTips = 10
tree = rtree(nTips)
x = rep(c("blue", "green"), each = 5)
names(x) = paste("t", 1:nTips, sep="")
plot(tree, tip.col=x[tree$tip.label])
nodelabels() # show you how your tree is indexed
# tiplabels()
intNodes = (Ntip(tree)+1L) : (Ntip(tree)+Nnode(tree))
sisters = Children(tree, intNodes)
# the next line assumes that the tree is strictly bifurcating, otherwise
its getting a bit more tricky
sisters = matrix(unlist(sisters), byrow=TRUE, ncol=2, dimnames =
list(intNodes, NULL))
# now the indices of sister clades are stored in the rows
desc = Descendants(tree) # descendants (tips) for each internal node (your
clades)
l = sapply(desc, length) # length of each clade
x = (x=="blue") # make x logical
x = x[tree$tip] # reorder your data to have the same order as the tree
l = sapply(desc, length) # length of each clade
blues = sapply(desc, function(d, x)sum(x[d]), x) # number of elements in a
clade, which are "blue"
greens = (l - blues) # number of elements in a clade, which are "green"
allgreen = (greens==l)
allblue = (blues==l)
# you can subset l, blues, greens, allgreen, allblue with the
res = rbind(cbind(greens[sisters[,1]],
blues[sisters[,2]])[allgreen[sisters[, 1]] & allblue[sisters[,2]], ],
cbind(greens[sisters[,2]], blues[sisters[,1]])[allgreen[sisters[, 2]]
& allblue[sisters[,1]], ])
(res = as.data.frame(res))
with this you should be able to compute your data.frame.
Cheers,
Klaus
On Thu, Oct 23, 2014 at 11:30 AM, Liam J. Revell <liam.rev...@umb.edu>
wrote:
> Hi Kevin.
>
> It sounds like what you want to do is perform a pre-order tree traversal
> and for each node visited ask if all the taxa to one side (e.g.,
> descendants of the right daughter node) are in state "0" and all the taxa
> to the other side (e.g., descendants of the left daughter) are in state
> "1". If this evaluates to be true, then you record the number of tips in
> each category. To do this most efficiently, you should not visit any
> daughters of a node for which you have evaluated balance - but if you do,
> then you will find that it doesn't satisfy the criterion described above
> (specifically, descendants on either side of the node will be all in state
> "0" or "1").
>
> This should be straightforward to code, but I do not have time to
> demonstrate right now. I will try to do it this evening. Let us know if you
> first figure it out yourself.
>
> - 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 10/23/2014 11:13 AM, Arbuckle, Kevin wrote:
>
>> Hi Fran�ois,
>>
>>
>>
>> Thanks again for your response. Wouldn't that lose the information of how
>> many species were in each clade? And how would I specify that 'clades' to
>> keep consist of those sharing either state? My original tree consists of
>> almost 3000 species so going through such clades manually would be
>> difficult at best, hence the need to automate it somehow. (I apologise, my
>> R coding skills are improving but still leave a lot to be desired in many
>> cases).
>>
>>
>>
>> I completely agree that BiSSE is far more appropriate for my aims, and
>> indeed this was the approach I used. However, reviewers have asked if I get
>> the same basic result using other methods, which is the only reason I am
>> attempting such analyses now.
>>
>>
>>
>> Thank you kindly once again for your time,
>>
>>
>>
>> Kev
>>
>>
>>
>> ________________________________
>> From: Fran�ois Michonneau [francois.michonn...@gmail.com]
>> Sent: 23 October 2014 16:05
>> To: Arbuckle, Kevin
>> Cc: r-sig-phylo@r-project.org
>> Subject: Re: [R-sig-phylo] Extracting sister groups
>>
>>
>> HI Kevin,
>>
>> If I understand correctly what you're trying to do, you'll first need
>> to collapse some of your tips to create clades, a proportion of which will
>> have the trait. You'll then be able to use this new tree to generate the
>> data.frame needed by the functions you mentioned in your original post.
>>
>> Depending on what you're trying to do, you may not want to lose this
>> phylogenetic information. Maybe a different approach, such as using BiSSE
>> in the diversitree package might be more appropriate?
>>
>> Cheers,
>> -- Fran�ois
>>
>> On Thu, Oct 23, 2014 at 10:27 AM, Arbuckle, Kevin <
>> k.arbuc...@liverpool.ac.uk<mailto:k.arbuc...@liverpool.ac.uk>> wrote:
>>
>> Hi Fran�ois,
>>
>> Thank you kindly for your offer of help. The code below will simulate a
>> phylogeny ("tree") and a dataframe ("trait") with one binary trait for 100
>> species. The format is representative of the data I am using for my
>> analyses so should serve as a test case. Hopefully this helps, let me know
>> if there's any other information I can provide.
>>
>>
>>
>> library(ape)
>> library(phytools)
>> tree<-rtree(100)
>> tran<-matrix(c(-1,1,1,-1),2,2)
>> rownames(tran)<-c("0","1")
>> colnames(tran)<-c("0","1")
>> phy<-sim.history(tree,tran)
>> trait<-data.frame(sp=tree$tip.label,bt=getStates(phy,type="tips"))
>> rownames(trait)<-tree$tip.label
>>
>> Cheers,
>>
>>
>>
>> Kev
>>
>>
>>
>> ________________________________
>> From: Fran�ois Michonneau [francois.michonn...@gmail.com<mailto:
>> francois.michonn...@gmail.com>]
>> Sent: 23 October 2014 14:54
>> To: Arbuckle, Kevin
>> Subject: Re: [R-sig-phylo] Extracting sister groups
>>
>>
>> Hi Kevin,
>>
>> We should be able to help you but it would be much easier if you
>> provided us with a small data set that illustrate the format of your
>> current dataset. How is your trait currently stored? and how is it
>> associated with the tips in your tree?
>>
>> Cheers,
>> -- Fran�ois
>>
>> On Thu, Oct 23, 2014 at 6:23 AM, Arbuckle, Kevin <
>> k.arbuc...@liverpool.ac.uk<mailto:k.arbuc...@liverpool.ac.uk>> wrote:
>> Hi everyone,
>>
>>
>>
>> I am attempting to run sister group analyses as one way to look at the
>> effect of a binary trait on diversification. Two of the functions from ape
>> that I'm looking at are diversity.contrast.test and richness.yule.test, but
>> both have the same limitation. They require the data to be input as a
>> dataframe of two columns, one with the number of species in clades that
>> have the trait of interest, and the other with the number of species in the
>> respective sister clades that don't have the trait. The issue is that I am
>> working with a very large tree, and so extracting and entering such
>> information by hand is not really feasible.
>>
>>
>>
>> I am therefore looking for a function which extracts all sister clades
>> that differ in the presence vs absence of the trait, and ideally is capable
>> of generating a dataframe of the appropriate format for the above functions
>> automatically. It seems that a function to do this should exist already,
>> but as I can't seem to find anything I would appreciate some help
>> (hopefully someone will know of such a function that already exists).
>>
>>
>>
>> Thanks,
>>
>>
>>
>> Kevin Arbuckle
>>
>> [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-phylo mailing list - R-sig-phylo@r-project.org<mailto:
>> 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-ph...@r-project.org/
>>
>>
>>
>> [[alternative HTML version deleted]]
>>
>>
>>
>> _______________________________________________
>> 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-ph...@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-ph...@r-project.org/
>
--
Klaus Schliep
Postdoctoral Fellow
Revell Lab, University of Massachusetts Boston
[[alternative HTML version deleted]]
_______________________________________________
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/
