[BioC] Removing genes from linear model fit advice sought

[Yong Li]

Dear Stephen,

I think a previous post by Gordon Smyth talking about filtering should 
answer most of your questions. You can find it at:

https://stat.ethz.ch/pipermail/bioconductor/2009-January/025827.html

Hope it helps.
Yong

stephen sefick wrote:
> I have done this with all of the genes in a microarray experiment
> including blanks, negative controls, empties, hk genes, and spike ins,
> genes of interest
> 1. RG <- backgroundCorrect(RG, method="normexp", offset=50)
> 
> 2. MA <- maNorm(as(RG, "marrayRaw"), norm="twoD")
> 
> 3. WA <- normalizeBetweenArrays(as(MA, "MAList"), method="scale")
> 
> This seems sensible to me.  Is it?
> 
> I am now thinking that I should remove everything except for the know
> differentially expressed genes and the genes of interest before
> fitting the linear model, contrasts, bayesian smotthing.  Is this a
> sensible coarse of action?  Thanks for all of your help in advance.
> kind regards,
>