[BioC] How to implement Stouffer's method using weights

[Ochsner, Scott A]

Hello,

I have found a number of published papers which have used Fisher's and Stouffer's methods for combining p-values in microarray meta analysis, however much of the discussion is naturally described as mathematical formulas.  As I'm not a statistician I'm having trouble interpreting these formulas as R
code.  The closest I have gotten is for Stouffer's: > pnorm(sum(qnorm(x))/sqrt(length(x))), which I got from a discussion on the BioC list serve archives.

I would like to implement these methods in my own meta analysis and need help with the code, especially with recommendations as to what type of experimental weights make sense and how to code them.  As a first thought I was thinking that the number of biological replicates would be a useful weight
to incorporate into the analysis.

Any help with a practical example of how to implement Fisher's and Stouffer's would be greatly appreciated.  
      

Scott A. Ochsner, Ph.D.
Molecular and Cellular Biology
Baylor College of Medicine
Houston, TX. 77030
phone: 713-798-6227