[BioC] basic limma questions
Gordon K Smyth
smyth at wehi.EDU.AU
Mon Mar 31 05:36:39 CEST 2008
I suspect the reason you didn't get a rsponse to your original posting is
that your questions are not very specific.
Assuming that you have three replicate arrays comparing a control to a
treatment, your analysis looks fine.
I can't give you any response to the problem reading ArrayVision format
because you haven't shown us the error you received or the structure of
your data files. Therefore we can't see if your files really are in
ArrayVision format or diagnose what the read problem is.
I can't tell you if the column names are correct, because ArrayVision
columns are used-defined. Your interpretation appears reasonable.
You ask if your design is correct, but you don't give any information
about your experiment, eg the targets information. I have to assume that
you have three replicate arrays.
In my opinion, there is no good way to combine irregular numbers of
within-array gene replicates, especially if the probes are not indentical.
Hope this helps
> Date: Fri, 28 Mar 2008 12:51:39 +0100
> From: "Endre Sebestyen" <endre.sebestyen at gmail.com>
> Subject: [BioC] basic limma questions
> To: Bioconductor <bioconductor at stat.math.ethz.ch>
> Again :
> I'm a beginner in limma and bioconductor, and I'd like to ask a few
> basic questions.
> I wrote the following script :
> targets <- readTargets("Targets1.txt")
> RG <- read.maimages(targets$FileName, columns=list(R="CtrlVol",
> G="DataVol", Rb="CtrlBg", Gb="DataBg"),
> RG$genes <- readGAL("maize.gal")
> RG$printer <- getLayout(RG$genes)
> spottypes <- readSpotTypes()
> RG$genes$Status <- controlStatus(spottypes, RG)
> bgCorr <- backgroundCorrect(RG, method="movingmin")
> nWithin <-normalizeWithinArrays(bgCorr, method="loess")
> nBetween <- normalizeBetweenArrays(nWithin, method="Aquantile")
> design <- c(1,1,1)
> isGene <- nBetween$genes$Status == "cDNA"
> fit <- lmFit(nBetween[isGene, ], design)
> fit <- eBayes(fit)
> res <- topTable(fit, number=1000)
> write.table(res, file="results24top1000.txt", sep="+++")
> First, limma didn't recognize the ArrayVision format, and I had to
> parse the raw data and define the columns myself. Is it correct to
> pass the CtrlVol to R, DataVol to G, etc? The other question is that
> I'm not sure about the design. Cy3 was the treated, Cy5 the control,
> but after I used the read.maimages function and defined the values
> myself, this design should be OK. Am I right?
> Last question : how can I combine the replicates on a chip? I have
> some genes with 2,3,etc replicates, but not all. This is the 46k maize
> array from www.maizearray.org
> Thanks for any comment and help.
> Endre Sebestyen
> Agricultural Research Institute of the Hungarian Academy of Sciences
> Department of Applied Genomics
> H-2462 Martonv?s?r, Brunszvik street 2.
More information about the Bioconductor