[BioC] another kooperberg question

michael watson (IAH-C) michael.watson at bbsrc.ac.uk
Wed Sep 14 19:17:40 CEST 2005


Hi Matt

Read in the data normally using:

RG <- read.maimages(etc etc)

Then read in the data *again*, using the commands I sent you a while ago ie:

read.series(etc) 
RGmodel <- kooperberg(etc)

IE do NOT pass RG to the kooperberg function.

The weights from RG should transfer easily across to RGmodel.

Mick

-----Original Message-----
From:	scholz at Ag.arizona.edu [mailto:scholz at Ag.arizona.edu]
Sent:	Wed 14/09/2005 11:13 AM
To:	michael watson (IAH-C); scholz at Ag.arizona.edu; bioconductor at stat.math.ethz.ch
Cc:	
Subject:	RE: [BioC] another kooperberg question

Thanks for your reply, Mick. Unfortunately, I've actually tried this before:

Rgmodel <- kooperberg(RG, fg="mean", bg="median", a=TRUE,layout=RGlayout)

and the problem is that I get the following error message:

"Error in matrix(0, ngenes, nslides) : cannot allocate vector of length 2105745408"

Again, I think the problem is that kooperberg doesn't accept an RG list as a
proper name argument. It would be nice if it did because there'd be no need to
transfer the weights (as they are already part of RG).

Matt


> Hi Matt
> 
> Hmmmm, something like this *should* work:
> 
> # read in the data normally just to get the weights
> RG <- read.maimages(targets$FileName, source="genepix", wt.fun=filters)
> 
> # read in the data using read.series and EXACTLT THE SAME DATA FILES as
> above
> # ...
> Rgmodel <- kooperberg(RG, fg="mean", bg="median", a=TRUE,
> layout=RGlayout)
> 
> # transfer the weights from RG to Rgmodel
> Rgmodel$weights <- RG$weights
> 
> 
> I think....
> 
> Mick
> 
> 
> -----Original Message-----
> From: bioconductor-bounces at stat.math.ethz.ch
> [mailto:bioconductor-bounces at stat.math.ethz.ch] On Behalf Of
> scholz at Ag.arizona.edu
> Sent: 12 September 2005 10:59
> To: bioconductor at stat.math.ethz.ch
> Subject: [BioC] another kooperberg question
> 
> Hi all,
> 
> I'm trying to run the Kooperberg correction on filtered data from
> genepix files.
> I've read my genepix data into and RGList (cleverly named RG) via a
> wt.fun function that filters out certain control spots that I don't want
> thrown into the analysis. In other words, I did:
> 
> RG <- read.maimages(targets$FileName, source="genepix", wt.fun=filters)
> 
> I now want to take this filtered data set and run kooperberg on it. The
> problem I'm having is that Kooperberg apparently requires a character
> vector containing genepix filenames as its initial argument but what I
> have is an RGList, not a character vector. In other words, I would like
> to do something like the following:
> 
> > kooperberg(RG, fg="mean", bg="median", a=TRUE, layout=RGlayout)
> 
> but RG is not an acceptable "names" argument. Obviously, I can't used
> the targets$FileName vector I used to initially read in the RGList
> because this is unfiltered data, and I can't use wt.fun within
> kooperberg. Any ideas on how to pass filtered data on to kooperberg?
> 
> Thanks
> 
> Matt
> scholz at ag.arizona.edu
>  
> 
> ---------------------------------------------
> College of Agriculture and Life Sciences Web Mail.
> http://ag.arizona.edu
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> 
> 


---------------------------------------------
College of Agriculture and Life Sciences Web Mail.
http://ag.arizona.edu



More information about the Bioconductor mailing list