[BioC] How to pool subgroups for makeContrasts() and subsequent limma analysis?

[James W. MacDonald]

Hi René,

On 1/23/2013 4:14 AM, René wrote:
> Dear all,
>
> My current analysis revolves around various groups and their 
> subgroups, which is why I would like to know how I can pool two or 
> more subgroups in a contrast matrix.
> As an example, let us assume I have 4 main groups (A, B, C and D), of 
> which one can be subdivided into 3 subgroups (B1, B2 and B3), hence
>
> groups = c("A","B1","B2","B3","C","D")
>
> Now, I would like to compare the combined/pooled group B (consisting 
> of B1, B2 and B3) against all other groups,
>
> B vs A, B vs C, B vs D.
>
> So my question is: how do I formalize that with makeContrasts() for an 
> analysis with limma?
> If it would be different matrices, I would just use cbind(), however I 
> am not sure whether B1+B2+B3 does the same thing in makeContrasts().

You are almost there. Note that you want the mean of the three groups, 
not the sum. So

makeContrasts((B1 + B2 + B3)/3 - A)

will e.g., do the comparison of B vs A.

Best,

Jim


>
> Thank you in advance and best regards
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: 
> http://news.gmane.org/gmane.science.biology.informatics.conductor

-- 
James W. MacDonald, M.S.
Biostatistician
University of Washington
Environmental and Occupational Health Sciences
4225 Roosevelt Way NE, # 100
Seattle WA 98105-6099