[BioC] lumi - limma contrast design

[Ryan C. Thompson]

I don't think that contrast answers the question as asked. The 
interaction contrast finds genes whose normal vs cancer ratio is 
affected by the drug. But Paolo is asking for genes that are 
significant in the C vs D contrast but not in the A vs B contrast. The 
solution, as has been discussed several times recently on this list, is 
to conduct two separate tests and do a set difference operation on them.

-Ryan

On Tue 15 Apr 2014 02:26:03 AM PDT, Axel Klenk wrote:
> Dear Paolo,
>
> sounds like you want:
> (D - C) - (B - A)
>
> Cheers,
>
>   - axel
>
>
>
> On Tue, Apr 15, 2014 at 9:16 AM, Paolo Kunderfranco <
> paolo.kunderfranco at gmail.com> wrote:
>
>> Dear All,
>>
>> Could you please help me to design a contrast for Differential Analysis?
>>
>> ....
>> constrast.matrix <- makeContrasts (
>> --------,
>> levels=design)
>> ...
>>
>>
>> I have 4 groups.
>>
>> A - normal
>> B - cancer
>> C - normal + drug
>> D - cancer + drug
>>
>> I would like to find genes modulated in D versus C. These genes should be
>> not affected or change significantly in B versus A
>>
>>
>> Any suggestion?
>>
>> Many thanks,
>>
>> Paolo
>>
>>          [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> 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
>>
>
>
>