R: Cluster Similarity Percentages With Inverted Y-Axis (Example Included)

Question

[Fucitol]

Hi All,

Although not really a bioinformatics question (only it's purpose in phylogenetics), I was wondering how I could perform a Bray Curtis similarity clustering in R in which I show the similarity percentages on an inverted Y-axis and all tree nodes ending at 100% as shown in the following picture (which I'm trying to replicate):

At the moment I create my plot in the following way (using S17 Bray Curtis dissimilarity measure, which just scales regular Bray Curtis to 0-100%):

library(vegan)
mat = 'some matrix'
d = (1 - vegdist(mat, method="bray")) * 100 
h = hclust(d)
plot(h)

Inverting the Y-axis (with ylim=c(100,80)) doesn't work. How can I create a dendogram as shown above from a distance matrix? Thanks for any help / advice!

As adviced, I've also posted this question at Cross Validated, see here

[David Quigley]

I don't have an answer for you, but you should also try the statistics wizards at http://crossvalidated.com (another Q&A site)

Answer 1

[Joseph Hughes]

13 months late but might still be useful:

library(vegan)
mat = mtcars
d = (1 - vegdist(mat, method="bray")) * 100
h = hclust(d)
plot(h, main = "Clustering cars using Bray Curtis method", sub = "", xlab="", axes = FALSE, hang = -1)
lines(x = c(0,0), y = c(0,100), type = "n") # force extension of y axis
axis(side = 2, at = seq(0,100,10), labels = seq(100,0,-10))

It looks like this:

[Marcus]

Just notice that

h = hclust(d)

[Joseph Hughes]

Thanks, I've corrected it.