[BioC] limma and marray data import problem

[Gordon K Smyth]

Dear Piotr,

The file extension "gpr" is short for GenePix Results file.  If ScanArray 
Express outputs a file with this extension, you should have every 
expectation that is formated exactly the same as a gpr file from GenePix, 
and therefore you should be able to read it using 
read.maimages(source="genepix").  If this is not true, then ScanArray is 
irresponsible to use this extension.

Same comments for the GAL file.  It is obviously not a GAL file as defined 
by GenePix, otherwise it would be read using readGAL().

>From your description below, a possible explanation for the problem is 
that your files have an extra column with no corresponding heading, e.g., 
a column of row numbers.  However no one on this mailing list can tell 
that for sure without you showing us some lines from your file.

Questions:
1. Why have you set row.names=NULL? This prevents R from detecting a 
column of row numbers. What happens if you remove this?

2. Are these files exactly as output by ScanArray, or have they been 
further processed?

3. Can you post the first few lines of an example file?

Best wishes
Gordon

PS. You posted the same question to the BioC mailing list on three 
consecutive days during the weekend.  Please post the question just once.


> Date: Sat, 31 May 2008 12:55:25 +0200
> From: " Piotr St?pniak " <piotrek.stepniak at gmail.com>
> Subject: [BioC] limma and marray data import problem
> To: bioconductor at stat.math.ethz.ch
>
> Hello Everyone,
>
> I am Piotr St?pniak, B.Sc. in Biotechnology, currently under M.Sc.
> course at Adam Mickiewicz University in Pozna?, Poland. I am working
> in Polish Science Academy in microarray experiments group.
>
> I'm a newbie in R and BioC, so please forgive me if my question is easy...
>
> I'm having problem with data import to RGList or marrayRaw objects.
> Using the following instruction:
> bialkoRaw<- read.maimages( dir(pattern="gpr"), columns=list(G="F543
> Median", Gb="B543 Median", R="F633 Median", Rb="B633 Median"),
> annotation=c("Block", "Column", "Row", "Name", "ID"), row.names=NULL)
> The data seems to load, but $genes table looks odd, I guess the column
> names are shifted right by 1 column:
> $genes
>  Block Column         Row Name   ID
> 1     1      1  ERG_Operon 2078 2647
> 2     2      1  ERG_Operon 2078 3102
> 3     3      1  ERG_Operon 2078 3549
> 4     4      1 FLT3_Operon 2322 3994
> 5     5      1 FLT3_Operon 2322 4444
> 2635 more rows ...
> This I think causes printer layout to be imported wrongly and then any
> other try to process the data (e.g. quality tests) produce such error
> message:
> Error in if (is.int(totalPlate)) { : argument is of length zero
>
> The data is obtained with ScanArrayExpress software, so I have it in
> gpr or csv files, both give similar errors, but loading csv files
> seems also to fail import values for each channel and gets only the
> file name headers.
>
> Marray import also fails, I will skip the info about it not to enlarge
> the mail unnecessarily.
>
> My R session info is as follows:
>> sessionInfo()
> R version 2.6.2 (2008-02-08)
> i486-pc-linux-gnu
>
> locale:
> C
>
> attached base packages:
> [1] grid      splines   tools     stats     graphics  grDevices utils
> [8] datasets  methods   base
>
> other attached packages:
> [1] arrayQuality_1.18.0  gridBase_0.4-3       hexbin_1.14.0
> [4] convert_1.16.0       RColorBrewer_1.0-2   cluster_1.11.10
> [7] arrayMagic_1.16.1    genefilter_1.16.0    survival_2.34-1
> [10] marray_1.18.0        vsn_3.6.0            limma_2.14.1
> [13] affy_1.16.0          preprocessCore_1.0.0 affyio_1.8.0
> [16] Biobase_1.16.3       lattice_0.17-7
>
> loaded via a namespace (and not attached):
> [1] AnnotationDbi_1.0.6 DBI_0.2-4           RSQLite_0.6-8
> [4] annotate_1.18.0     rcompgen_0.1-17
>
>
> I think I should also say that these data causes import problems to
> any other data analysis software :( I also tried to read the printer
> layout from gal file, but all I got was "Block, Row, Column, ID
> columns not found" error.
>
> I'd greatly appreciate any help, please.
>
> Yours faithfully,
> Piotr St?pniak