[BioC] Error with GEOquery

[Sean Davis]

On 4/19/06 6:28 AM, "Sharon" <sharonanandhi at gmail.com> wrote:

> Hi Sean,

Thanks for your help and for your effort in developing this useful
> tool. I
have tried now and it works fine. I was able to create the exprSet
> object
(followed vignette) from GSE. There are few things that I didn't
> understand.
I would appreciate your help in this.

1. When I looked at VALUE,
> it is mentioned that 'Log Ratio (not used)'. What
does this mean? BTW, this 2
> color cDNA data set.



All the information is supplied by the submitter (who in your case looks to
be Rishi Khan).  You will have to contact him/her about what information was
used/not used in their analysis.  The VALUE column typically represents
log-transformed, normalized values, but GEO doesn't enforce that in any
strict way.  Your best bet is to contact the original submitter to clarify
what data is meant to be useful.

> 2. Since, all the samples of from GPL2000, I tried to
> use getGEO("GPL2000")
to access gene information for this chip. But i don't
> know how to extract
this info from GPL. Could you please provide some  info on
> this issue.



Again, the platform information is supplied by the submitter.  Looking at
this GPL, it looks like the original submitter included only the "clone ID".
You will probably need to take these clone IDs and pass them through
something like:

http://source.stanford.edu

Just write out the clone id column from the GPL20000 by doing something
like:

gpl2000 <- GPLList(gse)[[1]] # the GPL information comes with the GSE--no
need to download again

write.table(Table(gpl2000)$CLONE_ID,file='rat.clone.id.txt',sep="\t",col.nam
es=FALSE,row.names=FALSE,quote=FALSE)

Then use the batch submission on source.stanford.edu to get the information
you want.  The output from stanford source will be a tab-delimited text file
that you can load back into R.

Hope that helps.

Sean