I am not sure you got my last mail so I sent it again,
Thanks, Anja
----Original Message-----
From: Gordon K Smyth [mailto:smyth@wehi.EDU.AU]
Sent: Tuesday, January 25, 2005 12:51 PM
To: Schiel, A.E. (HKG)
Cc: bioconductor@stat.math.ethz.ch
Subject: [BioC] vsn in limma or marray/arrayMagic
> Date: Tue, 25 Jan 2005 11:30:43 +0100
> From: <a.e.schiel@lumc.nl>
> Subject: [BioC] vsn in limma or marray/arrayMagic
> To: <bioconductor@stat.math.ethz.ch>
> Message-ID: <8A6C92C9EC1DD74AA66886D8DC3A38CE15545C@mailb.lumc.nl>
> Content-Type: text/plain
>
> Dear all,
> I do have a probably simple question, is running "vsn" in limma
(normalizeBetweenArrays) the same
> as running "vsn" in marray or arrayMagic?
Running "vsn" in limma is exactly the same as as running "vsn" in vsn,
in fact the vsn package is
simply used as a plug-in.
>>That's good to know, I was a bit concerned because you call vsn in
limma with the BetweenArray call instead of the WithinArray
>>command and I had the feeling that that call does normalize in a
different way. If it is ecaxtly the same than I can do my resorting of
>>genes with the MA object I create with the normalization and that
saves a lot of time!
As far as I know, you can't run vsn from marray. What do you mean?
>>I was actually running vsn from arrayMagic (sorry if I didn't make
that clear) and was just trying to figure out how to do that in marray
>>(but if you say it doesn't work than I can probably stop!)
If arrayMagic doesn't agree with limma and hence with vsn, then you
need to give precise code to
show the discrepancy. Perhap you have not used the code correctly.
>>For arrayMagic I used
>>description <- readpDataSlides(fileName, loadPath=loadPath,
deleteBlanks=TRUE, verbose=TRUE)
>>aD <- readIntensities(description, fileNameColumn ="fileName",
slideNameColumn="slideName", loadPath=loadPath,
>>+ type="GenePix")
>>norm <- normalise (aD, subtractBackground=TRUE, method="vsn")
>>in limma I used
>>RG <- read.maimages(targets$fileName, source="genepix")
>>RG$genes <- readGAL("sigmamousec_020603.gal")
>>RG$printer <- getLayout(RG$genes)
>>spottypes <- readSpotTypes("SpotTypes.txt")
>>RG$genes$Status <- controlStatus(spottypes, RG)
>>MA <- normalizeWithinArrays(RG)
>>MAvsn <-normalizeBetweenArrays(RG, method="vsn")
>>in both cases a defined targets before these functions. The Boxplot
generated with the quality-function in arrayMagic looks less "nice"
>>than the one I got in limma with
>>boxplot(MAvsn$M~col(MAvsn$M), names=colnames(MAvsn$M), main="vsn
data", col="darkgreen")
>>but in fact I do not know how the plots are actually generated in
arrayMagic!? Maybe the boxplot definition is not the default one.
> My Boxplots suggest that there is a difference and I was wondering
if someone can give me a hint
> why that might be.
> My particular problem is that I would like to use the fact that my
probes are spotted in
> duplicates but not the spacing limma expects. I therefore want to
first normalize all my arrays
> (22), then export the data to a text file and resort that file to
get the right spacing. If I am
> right I should be able to read such a text file with read.maimages
into limma and then continue
> with my analysis. Therefore I would like to know if I need to do my
normalization with marray (or
> arrayMagic) or if I can use vsn in limma?
I am puzzled why you would think there is a need to write out to a
text file at any stage.
>> Well I have sorting command lines by know which will transform my
data into the recommended form of duplicate
>>spots next to each other but I was not sure if I can use that on the
MAList objects in limma directly. But I tried this now and it worked
>>dataset <- (Ma)
>>ngenes <- nrow(dataset)
>>ncolumns <-ncol(dataset)
>>ind.gene <-rep(1:(ngenes/24),24)
>>ii.order <- order(ind.gene)
>>n.dataset <-dataset[ii.order,]
>>then continue with the limma analysis
>>Thanks, you remarks were very helpful (I am just starting with R so
I lack a few skills sorry for that!)
>> Anja
Gordon
> Many thanks in advance,
> Anja
Anja E. Schiel, Ph.D.
Departments of General Internal Medicine and Human Genetics
Leiden University Medical Center
PO Box 9503
2300 RA Leiden
The Netherlands
tel: -31-(0)71-5276067
fax: -31-(0)71-5276075
[[alternative HTML version deleted]]
