FDR analyses: minimum number of features

Dear Dr.  Graves

Many thanks for your response.  FDRs and their associated q values do
differ from Type I error rates and P values (See Storey and Tibshirani
PNAS 2003;100:9440-5).  It is an approach that is rapidly gaining
popularity in the analysis of genomic data where we have massive numbers
of covariates measured on a comparatively modest number of subjects.  To
my mind it is a real advance in dealing with the extreme multiple
comparisons problems that afflict such data.  Unfortunately, it is still
a relatively new technique and I do not believe that a consensus as to
the number of needed response features has been reached.

Submitting my question to the R-help list was a long shot, although
Storey's software for this methodology is written in R.

With best wishes,

Bill Dupont 

-----Original Message-----
From: Spencer Graves [mailto:spencer.graves at pdf.com]
Sent: Wednesday, September 21, 2005 9:27 AM
To: Dupont, William
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] FDR analyses: minimum number of features

	  Two thoughts on this:

	  1.  Your FDR (Not Franklin Delano Roosevelt) sounds like
another name for Type I error rate.  The definition of "reasonably
reliable FDRs" 
would seem to relate to the status of the literature on this issue among
researchers in genotyping.  As more reports of FRDs in genotyping are
published, I would expect that methodology for estimation and the
standard for accuracy would similarly evolve.

	  2.  Have you tried the Bioconductor (www.bioconductor.org/)
listserve?  They might be able to say something more useful than a
general list like this.

	  spencer graves