ClassDiscovery (version 3.4.0)

plotColoredClusters: Plot Dendrograms with Color-Coded Labels

Description

Provides an interface to the plot method for hclust that makes it easier to plot dendrograms with labels that are color-coded, usually to indicate the different levels of a factor.

Usage

plotColoredClusters(hd, labs, cols, cex = 0.8, main = "", line = 0, …)
pcc(hd, colors=NULL, …)

Arguments

hd

An object with S3 class hclust, as produced by the hclust function.

labs

A vector of character strings used to label the leaves in the dendrogram

cols

A vector of color names suitable for passing to the col argument of graphics routines.

cex

A numeric value; the character expansion parameter of par.

main

A character string; the plot title

line

An integer determining how far away to plot the labels; see mtext for details.

colors

A list; see details.

…

Any additional graphical parameters that can be supplied when plotting an hclust object.

Value

The function has no useful return value; it merely produces a plot.

Details

The plotColoredClusters function is used to implement the pltree methods of the Mosaic class and the PCanova class. It simply bundles a two step process (first plotting the dendrogram with no labels, followed by writing the labels in the right places with the desired colors) into a single unit.

The pcc function also produces dendrograms with colored annotations. However, instead of coloring the labels based on a single factor, it produces color bars for any number of factors. The colors argument should be a list with named components, where each component should correspond to a factor and a color scheme. Specifically, the components must themselves be lists with two components named fac (and containing the factor) and col (containing a named vector specifying colors for each level of the factor).

See Also

hclust, Mosaic, PCanova, par

Examples

# NOT RUN {
# simulate data from three different groups
d1 <- matrix(rnorm(100*10, rnorm(100, 0.5)), nrow=100, ncol=10, byrow=FALSE)
d2 <- matrix(rnorm(100*10, rnorm(100, 0.5)), nrow=100, ncol=10, byrow=FALSE)
d3 <- matrix(rnorm(100*10, rnorm(100, 0.5)), nrow=100, ncol=10, byrow=FALSE)
dd <- cbind(d1, d2, d3)

# perform hierarchical clustering using correlation
hc <- hclust(distanceMatrix(dd, 'pearson'), method='average')
cols <- rep(c('red', 'green', 'blue'), each=10)
labs <- paste('X', 1:30, sep='')

# plot the dendrogram with color-coded groups
plotColoredClusters(hc, labs=labs, cols=cols)

# simulate another dataset
fakedata <- matrix(rnorm(200*30), ncol=30)
colnames(fakedata) <- paste("P", 1:30, sep='')
# define two basic factors, with colors
faccol <- list(fac=factor(rep(c("A", "B"), each=15)),
               col=c(A='red', B='green'))
fac2col <- list(fac=factor(rep(c("X", "Y", "Z"), times=10)),
               col=c(X='cyan', Y='orange', Z='purple'))
# add another factor that reverses the colors
BA <- faccol
BA$col <- c(A='blue', B='yellow')
# assemble the list of factors
colors <- list(AB=faccol, XYZ=fac2col, "tricky long name"=fac2col,
               another=BA)
# cluster the samples
hc <- hclust(distanceMatrix(fakedata, "pearson"), "ward")
# plot the results
pcc(hc, colors)


#cleanup
rm(d1, d2, d3, dd, hc, cols, labs)
rm(fakedata, faccol, fac2col, BA, colors)
# }