[BioC] liimma and Across Array Normalisation

[Steve Lianoglou]

Hi Saket,

On Tue, Feb 11, 2014 at 3:49 PM, Saket Choudhary <saketkc at gmail.com> wrote:
> I wanted it to visualise it in 3x2 style.

Whipping up your own MA plot isn't very difficult from the raw data.

Given a set of measurements A from the transcriptome, and another set
of measurements B from the transcriptome, the x,y coordinate for each
gene_i is:

* x: mean expression of gene_i between A and B, ie: (A_i + B_i) / 2
* y: fold change of gene_i A vs B, ie: A_i / B_i

(vectorize these calculations and be sure to convert to logs where
appropriate for your data -- that is to say, if it's not log already)

Given that you already have the data on hand that you want to plot
(data from several single channel arrays), and the details of
constructing a basic MA plot are quite straightforward, you can take
your data + your favorite plotting package (base, lattice, ggplot2)
and plot it in any number of ways that you see fit.

It'd be an informative exercise and ultimately give you the most
flexibility in the data presentation, which seems like what you are
after, so perhaps you can try that.

HTH,
-steve

-- 
Steve Lianoglou
Computational Biologist
Genentech