[BioC] Limma: nestedF and interaction term

[James W. MacDonald]

Hi Massimo,

On 9/30/2010 10:54 AM, MassimoAcquaviva at ospedale-gaslini.ge.it wrote:
> Hallo to everyone,
>
> In have a  model  like the factorial design example in the limma user
> guide
>
> TS<- factor(TS, levels = c("WT.U", "WT.S", "Mu.U", "Mu.S"))
> design<- model.matrix(~0 + TS)
> colnames(design)<- levels(TS)
> cont.matrix<- makeContrasts(
>   WT.SvsU = WT.S - WT.U,
>   Mu.SvsU = Mu.S - Mu.U,
>   Diff = (Mu.S - Mu.U) - (WT.S - WT.U),levels = design)
> fit<- lmFit(eset, design)
> fit2<- contrasts.fit(fit, cont.matrix)
> fit2<- eBayes(fit2)
>
> I am planning to extract the common response to the treatment between  WT
> and Mu  using nestedF
>
> cont.matrix<- makeContrasts(
>   WT.SvsU = WT.S - WT.U,
>   Mu.SvsU = Mu.S - Mu.U)
> fit<- lmFit(eset, design)
> fit2<- contrasts.fit(fit, cont.matrix)
> fit2<- eBayes(fit2)
> results<- decideTests(fit2,method="nestedF",p=0.05)
>
> and to extract the different response testing the interaction term with
> moderated t statistics
>
> cont.matrix<- makeContrasts(
>    Diff = (Mu.S - Mu.U) - (WT.S - WT.U),levels = design)
> fit<- lmFit(eset, design)
> fit2<- contrasts.fit(fit, cont.matrix)
> fit2<- eBayes(fit2)
> toptable(fit2,adjust="BH")
>
> and finally make a Venn Diagram with the results of the three comparisons
>
> Is this approach correct to determinate which genes respond only in WT,
> which ones only in Mu and which, among the common ones, are differently
> expressed?

I would tend to keep all comparisons in the contrasts matrix and use 
decideTests() to sort out the significantly different contrasts, rather 
than splitting the interaction out separately.

Note that this parameterization is different from the classical 
parameterization for factorial designs, so there isn't IMO any reason to 
preclude all contrasts from being significant simultaneously.

Best,

Jim


>
> I would really appreciate some help
>
> Thank you in advance
>
> Massimo Acquaviva
>
> Department of Molecular Biology
> Giannina Gaslini Institute, Genova Italy
>
>
>
>
>
>
>
>
>
>
>
>
>
> Aiutaci ad aiutare: Il tuo 5 per mille a sostegno della RICERCA.
> Come fare: nella prossima dichiarazione dei redditi ti chiediamo di
> apporre la firma nel riquadro "Finanziamento agli enti della ricerca
> sanitaria"
> e il codice fiscale dell'Istituto Giannina Gaslini: c.f. 00577500101.
> Istituto Giannina Gaslini - www.gaslini.org
> .
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> 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
Douglas Lab
University of Michigan
Department of Human Genetics
5912 Buhl
1241 E. Catherine St.
Ann Arbor MI 48109-5618
734-615-7826
**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues