conditional coloring of image labels

On Friday 11 November 2005 10:37 am, Jake Michaelson wrote:

I'll clarify a little and hopefully this will make more sense (thanks for
the
friendly encouragement):

Let's say I have 6 samples and am looking at 3 genes, with intensities in
a

matrix as follows:

genes=cbind(ABC1=c(3,4,4,5,6,3), ABC2=c(4,3,4,7,7,8),

ABC3=c(8,7,8,6,3,2))

genes

ABC1 ABC2 ABC3
[1,] 3 4 8
[2,] 4 3 7
[3,] 4 4 8
[4,] 5 7 6
[5,] 6 7 3
[6,] 3 8 2

###plot the image

image(1:nrow(genes), 1:ncol(genes), genes, axes = FALSE, xlab = "", ylab

=
"", col=cm.colors(256))

###label the axes

axis(1, 1:nrow(genes), labels = rownames(genes), las = 2, line = -0.5,

tick =
0)

axis(2, 1:ncol(genes), labels = colnames(genes), las = 2, line = -0.5,

tick =
0)

Now let's say (I'm just making these numbers and the scenario up here --
for
simplicity's sake) that I had run a statistical analysis previously and
wanted to label the genes that showed a significance of p< 0.05. Let's say
that ABC1 and ABC3 had p<0.05 (assume that these values would be in a
two-column matrix with the gene name and its p-value).

sig=cbind(name=c("ABC1", "ABC2", "ABC3"), pvalue=c(0.005, 0.1, 0.001))
sig

name pvalue
[1,] "ABC1" "0.005"
[2,] "ABC2" "0.1"
[3,] "ABC3" "0.001"

I now want these (the names of the significant genes) to be labeled in red
rather than black on the plot. I would eventually write a script that
would
generate a large number of these images, each with a different set of
genes.
I would like to insert some sort of conditional formatting so that if that
gene meets the significance threshold, the name is automatically plotted
in
red on the plot.

I hope this is more clear and effective in explaining what I'm looking
for.


Thanks,

--Jake

Hi all,

I am interested in plotting a heatmap of a set of genes. I would like

the

text labels of these genes to be colored red rather than black if they

meet

certain statistical criteria (using an if statement). I'm not sure how

to

change individual color labels without changing them all. Can anyone
provide some insight on how to do this?

Thanks in advance,

Jake