Hi Kamila,
sounds like you want to treat the integers categorically, in which case for
neighbour joining you could try generating a distance matrix using Gower
dissimilarity (e.g. using function 'daisy' in Cluster). If you specify that the
data are factors, the function automatically treats them as nominal variables.
> require(cluster)
> DATA <- matrix(data=c(0,1,2,3,3,1,2,2,4,3,2,2,5,5,4,1,2,5,5,5), nrow=4,
> ncol=5, byrow=TRUE)
> rownames(DATA) <- c("species1", "species2", "species3", "species4")
> colnames(DATA) <- c("f1", "f2", "f3", "f4", "f5")
> DATA
f1 f2 f3 f4 f5
species1 0 1 2 3 3
species2 1 2 2 4 3
species3 2 2 5 5 4
species4 1 2 5 5 5
> FUNCx <- function(x) as.factor(x)
> DATA2 <- as.data.frame(apply(DATA, 2, FUNCx))
> DAISY <- daisy(DATA2)
> DAISY
Dissimilarities :
species1 species2 species3
species2 0.6
species3 1.0 0.8
species4 1.0 0.6 0.4
Metric : mixed ; Types = N, N, N, N, N
Number of objects : 4
> require(ape)
> plot(nj(DAISY))
Cheers,
Greg Guerin
----------------------------------------------------------------------
Message: 1
Date: Wed, 29 May 2013 15:21:07 +0000
From: "Naxerova, Kamila" <[email protected]>
To: "[email protected]" <[email protected]>
Subject: [R-sig-phylo] "equal or not" distance function for NJ
Message-ID:
<[email protected]>
Content-Type: text/plain; charset="us-ascii"
Dear list,
I would like to construct a neighbor-joining tree from a data set in which each
taxon is characterized by a vector of integers ranging from 0 to 5. I would
like to treat these as unordered characters, i.e. assuming that the likelihood
of transitioning from 0 to 1 is the same as transitioning from 0 to 5. For
this, I would need some kind of "equal or not" distance function which
increases the distance between two taxa by one for each dimension that differs,
regardless of its magnitude.
Is such a distance function already implemented anywhere or is it easiest to
create one for myself?
Many thanks.
Kamila
------------------------------
_______________________________________________
R-sig-phylo mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
End of R-sig-phylo Digest, Vol 64, Issue 20
*******************************************
_______________________________________________
R-sig-phylo mailing list - [email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-phylo
Searchable archive at http://www.mail-archive.com/[email protected]/
