[BioC] LIMMA question

[Gordon K Smyth]

This is a repeated measures time course experiment.  You should consult a statistician over this. 
In limma, the possibilities are to treat each subject as a block (see duplicateCorrelation etc) or
to fit subjects as fixed effects and compare times within subjects.  Either of these approaches
may work fine for you, and are easy to apply, but it would be best to get some advice from someone
who understands repeated measures and can guide you.

Gordon

> Date: Thu, 12 May 2005 08:26:36 -0500
> From: "Richardson, Dawn K" <richardsond2 at uthscsa.edu>
> Subject: [BioC] LIMMA question
> To: <bioconductor at stat.math.ethz.ch>
>
> Dear All,
>
> My experimental set up is as follows:
>
> Time course experiment in 10 subjects each given treatment for 240
> minutes.
>
> During the course of the treatment I took a muscle biopsy at time 0, 30
> and 240 min. I extracted RNA from the biopsies at all time points and
> used the Affymetrix HG-U133A chips for each sample.
>
> I normalized all CEL files together using GC-RMA and I'm at the stage
> where I want to look for genes that are differentially expressed at all
> time points; i.e. time 0 versus time 30, time 0 versus time 240, and
> time 30 versus 240.
>
> I have looked in the LIMMA manual and I'm not sure which script I need
> to use to perform this test.
>
> Any help or suggestions would be appreciated,
>
> Thanks,
>
> Dawn.