This file demonstrates how to construct a colorimetric comparision array (CCA)
using the Matlab software package, which runs on Unix and Windows systems.  We
will generate the CCAs published in Shamji et al., "Partitioning the transcriptional
program induced by rapamycin among the effectors of the Tor proteins" using the 
actual data.

To actually do these examples, you must be running Matlab and be in the directory 
with (or have path access to) the other files from the above publication's 
supplementary material.

Load the two files of data, angles.txt and magnitudes.txt, (which are merely 
tab- and carriage-delimited matrices of numbers) by typing:

>> load('angles.txt');
>> load('magnitudes.txt');

Now we will load the descriptors for the experiments (top row) and the gene sets 
(right column).  These files are carriage-return delimited text lists.  The function 
castringload loads these lists into Matlab cell arrays.

>> titles=castringload('titles.txt');
>> subsets=castringload('subsets.txt');

It is simple to see the contents of, for example, subsets.  Simply type the 
variable name.

>> subsets

subsets = 

  Columns 1 through 4

    'whole genome'    'NDP'    'storage carb'    'hexose transporters'

  Columns 5 through 8

    'DNA replication'    'purine biosynth'    'proteasome'    'TCA cycle'

  Columns 9 through 12

    'oxidative phos'    'ubiquinone synth'    'ATP synth'    'RNA Pol I'

  Columns 13 through 16

    'RNA Pol II'    'RNA Pol III'    'tRNA processing'    'tRNA synthet'

  Columns 17 through 19

    'mito tRNA synthet'    'r-proteins'    'mito r-proteins'

  Columns 20 through 22

    'ribosomal biogen'    'rRNA processing'    'translation init'

  Columns 23 through 25

    'translation elong'    'mRNA export'    'mRNA decay'


To generate a CCA, invoke the following command:

>> angleccamake(angles,180,90,0,titles,subsets,'wt1 vs wt1 + rap');

A CCA should appear on the screen (though it may not be correctly formatted).  The 
syntax of the angleccmake is as follows:


	Argument1 = the matrix of data to be visualized
	Argument2 = the numerical value which is assigned the color green
	Argument3 = the numerical value which is assigned the color yellow
	Argument4 = the numerical value which is assigned the color red
	Argument5 = a cell array of strings containing the column headings
	Argument6 = a cell array of strings containing the row names
	Argument7 = a string describing the reference profile 

Note that dimensions must agree.  For example if the # of rows in Argument1 exceeds 
the # of strings in Argument6, you will get an error.  Note that you can assign any 
numbers you want to green and red, but yellow must lie somewhere in the middle.

To generate a ratio of vector magnitudes CCA, type:

>> rvmccamake(magnitudes,0,1,2,titles,subsets,'wt1 vs wt1 + rap');

There is fundamentally no difference between angleccamake or rvmccamake.  They are 
customized variants to produce slightly different legends.  In principle, a single 
function ccamake could be written with more arguments that accomplishes both tasks.  
Note that in this example, the green value is less than the red value, unlike before.  
It is completely user-defined.

All of these M-files that are invoked above are very simple programs that you can 
easily understand or modify to suit your own purposes.  Alternative color schemes can 
be created by modifying (or swapping in a new version of) the file "profilecolors.m" 
which is called by both angleccamake and rvmccamake.  This file executes the task of 
generating the RGB vectors which are the spot colors.  With regard to formatting, 
there are parameters in the first lines of rvmccamake.m and angleccamake.m that can be 
modified for a particular design.  These particular settings were designed for 
printing, not for on-screen visualization. These CCAs can be exported to postscript 
files within Matlab and visualized inside programs like Ghostscript.