[BioC] Getting additional columns from Imagene files with read.maimages from limma

[Gordon K Smyth]

Dear Peder,

The first issue you raise is not a problem.  You can access components 
whose names contain spaces simply by quoting the names:

   RG$other[["G Signal Stdev"]]
   RG$other[[1]]
   RG$other$"G Signal Stdev"

The second issue, assigning rownames to an RGList object with 'other' 
columns, is a bug.  I've committed a fix to Bioconductor in limma 2.16.5 
this morning.  You should be able to reinstall the new version of limma in 
a couple of days.

Best wishes
Gordon

> Date: Wed, 18 Mar 2009 13:48:06 +0100
> From: "Peder Worning" <pwo at exiqon.com>
> Subject: [BioC] Getting additional columns from Imagene files with
> 	read.maimages from limma
> To: "bioconductor" <bioconductor at stat.math.ethz.ch>
>
> Hello Bioconductor,
>
> I am using the read.maimages function from the limma package. But I want
> to read two more columns from the Imagene files than default. I have
> tried to use the other.columns variable to read the data and it works in
> principle, I get the data but not in a format I can use and the "other"
> option spoils my naming routine for the spots in the arrays. I'll show
> some code and ask how to get around it. In this experiment I have 3
> slides with 14592 spots
>
> This is how I normally do it and each spot gets its id as the rowname
> RG.NQ <- read.maimages(file.NQ, source="imagene",
>        names=target.NQ$Barcode,
>        wt.fun=mywtfun(c(1,2,3,4,5,6,7)),
>        columns=list(f="Signal Mean",b="Background Median"))
>
> rownames(RG.NQ) <- RG.NQ$genes[,6]
> dim(RG.NQ)
> [1] 14592     3
>
> names(RG.NQ)
> [1] "R"                "G"                "Rb"               "Gb"
>
> [5] "source"           "Field.Dimensions" "weights"          "printer"
>
> [9] "genes"
>
>
>
> Here follow the new code that gives me troubles
> RGp.NQ <- read.maimages(filep.NQ, source="imagene",
>        names=targetp.NQ$Barcode,
>        wt.fun=mywtfun(c(1,2,3,4,5,6,7)),
>        columns=list(f="Signal Mean",b="Background Median"),
>        other.columns=list(t="Signal Stdev",u="Background Stdev"))
> names(RGp.NQ)
> [1] "R"                "G"                "Rb"               "Gb"
>
> [5] "source"           "Field.Dimensions" "weights"          "other"
>
> [9] "printer"          "genes"
> names(RGp.NQ$other)
> [1] "G Signal Stdev"     "G Background Stdev" "R Signal Stdev"
> [4] "R Background Stdev"
>
> And here starts the problems. How do I read variable names with white
> spaces \ or \\ do not work, and how do I recover the original format of
> the RGp.NQ object from the list RGp.NQ$other
> dim(RGp.NQ)
> [1] 14592     3
> 3 slides 14592 spots
>
> Furthermore I cannot make each spot have its id as rowname
>
> rownames(RGp.NQ) <- RGp.NQ$genes[,6]
> Error in dimnames(object$other[[a]]) <- value : object "object" not
> found
>
> If anyone knows a way around these troubles I'll appreciate it a lot
> because I need the Signal Stdev and the Background Stdev to evaluate the
> quality of each spot.
>
> Regards
> Peder Worning
>
>
> Best regards
>
> Exiqon A/S
>
> Peder Worning, Ph.D.
>
> Senior Scientist, Biomarker Discovery
>
> Telephone: +45 45650457
> Telefax: +45 45661888
> E-mail: pwo at exiqon.com
>
> Bygstubben 3
> DK-2950 Vedbaek
> Denmark
> www.exiqon.com