[BioC] Limma Design

[J.delasHeras at ed.ac.uk]

Quoting "claudio.is at libero.it" <claudio.is at libero.it>:

> Dear BioC,
>
> I am tring to use Limma package in order to analyze miRNA   
> expression. I would like to identify differentially expressed miRNA;  
>  my experimental design encompasses dye swap, and different  
> technical  replicates:
>
> REF expA1
> REF expB1
> REF expC1
> the corensponing dye-swap
> expA1 REF
> expB1 REF
> expC1 REF
>
> REF expA2
> REF expB2
> REF expC2
> the corensponing dye-swap
> expA2 REF
> expB2 REF
> expC2 REF
>
> REF expA3
> REF expB3
> REF expC3
> the corensponing dye-swap
> expA3 REF
> expB3 REF
> expC3 REF
>
> I how can I set the design matrix for linear model in order to   
> identify differentially regulated genes in respect to the   
> comaparison between A vs B, A vs C and B vs C. I tried with
> fit <- lmFit(MA, design=c(1,2,3,-1,-2,-3,1,2,3,-1,-2,-3,1,2,3,-1,-2,-3))
>
> but I get results that are too much different for t-test among   
> conditions, so I think it is wrong.
>
> thank you
>
> Claudio

Hi Claudio,

take a look at the Limma users guide, Chapter 7. It tells you how to  
create a design and contrasts matrix, etc.
Functions like 'model.matrix', 'makeContrasts' and 'contrasts.fit'  
will make the process quite straightforward.

Jose

-- 
Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131 6513374
Institute for Cell & Molecular Biology        Fax:   +44 (0)131 6507360
Swann Building, Mayfield Road
University of Edinburgh
Edinburgh EH9 3JR
UK