[BioC] creating an MA plot to compare treatments [was: creating an MA plot with two single channel agilent arrays]

Gordon K Smyth smyth at wehi.EDU.AU
Wed Jul 6 01:47:50 CEST 2011

Dear Yifang,

Firstly, if you have single channel data, please read it into limma as an
EList, not as an RGList object.

The plot you are trying to create is called a fitted model MA-plot in
limma.  You get it by first fitting the linear model to compare your two
treatments, then making the MA-plot.  Suppose that y is your EList object.
You might use:

design <- model.matrix(~condition)
fit <- lmFit(y, design)
plotMA(fit,array=2)

If condition has two levels, then this will plot the logFC between your
two treatments versus the mean A-value of your arrays.  If you have two
groups with equal numbers of arrays, then this is the same as the MA plot
of mean-values that you proposed.  The fitted model MA-plot however is
much more general, and can plot any treatment comparison.  The 'array'
argument specifies in this context which coefficient of the linear model
you want to plot.

Best wishes
Gordon

---------------------------------------------
Professor Gordon K Smyth,
Bioinformatics Division,
Walter and Eliza Hall Institute of Medical Research,
1G Royal Parade, Parkville, Vic 3052, Australia.
Tel: (03) 9345 2326, Fax (03) 9347 0852,
smyth at wehi.edu.au
http://www.wehi.edu.au
http://www.statsci.org/smyth

On Tue, 5 Jul 2011, yifang tan wrote:

>
> Hi, Gordon!
>
> MA-plot the mean values of two treatments. Say: I have RG\$G as EList,
> RG\$G[, c(1,5,9,13)] as condition-1, RG\$G[, c(2,6,10,14)] as condition-2,
> I want to MA-plot of the mean values for each condition. The reason I do
> this is that my platform is single channel microarray, and the
> individual array looks fine but I thought checking the mean value would
> be another angle to look at the data, which seems reasonable too. (I am
> not sure, though!)

> I tried the regular plotting, which works fine:
> plot(apply(RG\$G[, c(1,5,9,13)], 1, mean), apply(RG\$G[, c(2,6,10,14)], 1, mean))
>
> How could I accomplish the plotMA of the two means? The trick maybe to create the EList of the two mean values, and how?
>
> Thanks!
>
> Yifang
>
>
> Yifang Tan

______________________________________________________________________
The information in this email is confidential and intend...{{dropped:4}}