[BioC] Re: Just a quick question

[Gordon K Smyth]

> Gordon,
>
> Just a quick question using Limma.  I want to form a table with p-value
> in columns and genes in rows for multiple contrasts and for all genes.

It seems to me that the eBayes() function does exactly this.

> Ultimately, I want to be able to determine the genes that have p<alpha
> for some proportion of the experiments (2/3, for example).

I don't believe in this sort of analysis.  One combined p-value seems
better.  But the p.value component returned by eBayes() makes it easy if
you want to do it.

Gordon

> A more involved question that you needn't answer if you don't have time,
> but could I use classifyTests to do the same thing (find all genes that
> are "significant" over a set of experiments)?
>
> Thanks,
> Sean