[BioC] DEGraph demo exprData question

[Pierre Neuvial]

Hi Dick,

>
>
> Hi,
>
> Does anyone know how the values in the DEGraph demo exprData matrix were
> calculated?  I have read the DEGraph documentation, as well as the source
> documents for the data (GSE6532), and I can't figure out how to go from the
> RMA values (range 0-16) to the values in the exprData matrix (-5 to 7.5
> about).  It doesn't seem to be z-scores.
>
> If anyone can help, I'd love to hear.  Just trying to understand the details
> of DEGraph

Thanks for your interest in DEGraph.  The (probe set) expression
values from the demo were taken from the "LUMINAL.RData.gz" file in
the GSE6532 GEO archive:

ftp://ftp.ncbi.nih.gov/pub/geo/DATA/supplementary/series/GSE6532/LUMINAL.RData.gz

These files were processed by RMA, and *median-centered*, as explained
in the corresponding README file
(ftp://ftp.ncbi.nih.gov/pub/geo/DATA/supplementary/series/GSE6532/LUMINAL_README.txt.gz):

"RMA normalization was performed separately for each population
("series" in demo) and all the probesets were median centered."

Note that 'population' refers to the fact that the data comes from
three different institutions.

This median-centering explains the range of expression values in
"LUMINAL.RData", and therefore the range of the corresponding values
in the demo data set of the DEGraph package.

In case you'd like more information on what exactly is in DEGraph's
'exprData': the data in GSE6532 is at the *probe set* level and the
data in DEGraph is at the *gene* level.  We defined the expression
level of a gene as the expression level of the probe set with largest
alignment score among all probe sets mapping to this gene according to
the annotation in GSE6532.  When the largest alignment score was
achieved by several probe sets, we took the median expression level of
those probe sets.  I've just posted the script we have been using to
do this here:

http://stat.genopole.cnrs.fr/_media/Members/pneuvial/Loi2008_geneLevelData.R

I hope this helps,

Pierre

>
> Thanks very much,
> Dick
> *******************************************************************************
> Richard P. Beyer, Ph.D. University of Washington
> Tel.:(206) 616 7378     Env. & Occ. Health Sci. , Box 354695
> Fax: (206) 685 4696     4225 Roosevelt Way NE, # 100
>                        Seattle, WA 98105-6099
> http://depts.washington.edu/ceeh/members_fc_bioinfo.html
> http://staff.washington.edu/~dbeyer
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives:
> http://news.gmane.org/gmane.science.biology.informatics.conductor