[BioC] limma topTable question

[Jenny Bryan]

On Apr 5, 2005, at 4:37 PM, Gordon Smyth wrote:

 > I am not convinced that this would be a useful option. Microarray
 > differential expression analyses are virtually always two-sided, for
 > good reason, because researchers need to know about genes moving
 > strongly down as well as up. Hence topTable() is designed to
 > facilitate a two-sided analysis. You are implying that you want to do
 > an anlaysis in which you don't want to even see the genes moving down.
 > Why do you think that this is a generally useful analysis? I haven't
 > seen any microarray problems which I would want to analyse that way.

Although my own thinking is the same as Gordon's w.r.t. two-sided 
analyses, I have actually noticed this phenomenon several times in 
collaboration.  That is, a much greater interest in up-regulated genes 
than down-regulated genes.  I gather there are at least two reasonably 
good explanations for this:

1) If the next step is some sort of intervention aimed at changing 
transcriptional activity, it is easier and/or more likely to be an 
attempt to decrease or completely suppress transcription than to 
increase it.  Hence, the preference for finding overexpressed genes.

2)  First, take it as given that the microarray technology only has 
hope (under the usual protocols for preparing the mRNA) of measuring 
expression changes in a gene's *share* of a typical mRNA pool, not 
absolute changes on a copies per cell basis.  Then, to construct a 
highly artificial example that makes the point, imagine that the 
treatment induces substantially higher absolute expression of some 
genes, but leaves the expression of most everything else practically 
the same (this *is* on an absolute, copies per cell basis).  Maybe some 
type of stress response is being provoked and studied.  Then, the true 
relative share fold-changes we hope to estimate with array data will 
evidence a mix of up and down regulation.  But we only care about 
identifying the genes that are up-regulated.

I make no claims about these one-sided analyses being generally 
advisable, but since it does come up, these are my attempt to 
understand why.

Jenny Bryan