Hello.

This directory includes prototype code to
enable some of the major calculations described in 

Newton, M.A., Sarkar, D., Noueiry, A., and Ahlquist, P. (2003).
Detecting differential gene expression with a semiparametric hierarchical
mixture model.  Technical Report 1074-REVISED.  Department of Statistics,
University of Wisconsin Madison.  DECEMBER 2003.

Under consideration is the comparison problem when
we have replicated microarrays in each of two conditions.
The methods compute posterior probabilities of differential
gene expression using a semiparametric hierarchical mixture model.
See www.stat.wisc.edu/~newton/papers/abstracts/tr1074a.html

The code is for R, version 1.7.1 or later (see www.r-project.org)
and is subject to the GNU Public License version 2.0 or later.

Although it would be beneficial in some respects, I have
chosen not to compile the functions into an R library.  It is
the longer term goal of the research project to extend the
EBarrays R library to include the semiparametric methodology introduced
here (Kendziorski et al, 2002).  For now, I provide some stand-alone R 
functions that you may try for the two-sample comparison of normalized 
expression measurements.

Contents:

cvdiag.R               Code to make diagnostic plots for checking 
                       the distribution of expression conditional on means.

spfit.R                Main estimation routine for fitting semiparametric
                       model.  

example.R              Source code to demonstrate the above functions.
                       Data are not provided here.


Please send comments to newton@stat.wisc.edu.

Cheers,

Michael Newton,
Madison, WI. December, 2003.