[BioC] Limma: How to read gene list , coordinates of sport when NO GAL file available

[J.delasHeras at ed.ac.uk]

Quoting Srinivas Iyyer <srini_iyyer_bio at yahoo.com>:

> Dera group,
> limma is an excellent module for gene expression data
> preprocessing and analysis.
> however, I looked into many places i did not find a
> good tutorial when the .gpr file is not what I it
> should look like. Also, when GAL file is not the same
> what it should be.
>
> I have a dataset downloaded from ArrayExpress and has
> the following column names:
>
> [B635+1SD	B635+2SD	Autoflag	B Pixels	B635	B635 CV	B635
> Mean	B635 Median	B635 SD	Circularity	Dia.	F Pixels
> F635 % Sat.	F635 CV	F635 Mean	F635 Mean - B635	F635
> Median	F635 Median - B635	F635 SD	F635 Total Intensity
> Flags	Normalize	SNR 635]
>
>
> The chip definition file obtained from "Array design
> used" section of ArrayExpress has the following
> columns:
>
> [MetaColumn	MetaRow	Column	Row	Reporter Identifier
> Reporter Name	Reporter Biosequence Type	Reporter
> actual Sequence	Reporter Comment	Reporter Group Role
> Reporter Control Type	CompositeSequence Identifier
> CompositeSequence Name	Composite Sequence Comment]
>
> when i did:
> dat <- read.maimages('filename',source
> ='genepix.custom')
>
> I get "Error in readGPRHeader(fullname) : File is not
> in Axon Text File (ATF) format"
>
>
> my questions are:
>
> what should I tell read.maimages to accept my file and
> process further.
>
> what should I do when I do not have GAL file.  how can
> the other file help me get genelist etc.
>
> Please help me.
>
> Thanks
> sri

I am not sure what's the simplest way to go about it, but when in 
doubt, I'd just define everything myself.

I'd probably make my own GAL file first. You need to convert the MR MC 
R C coordinates you have into Block, Row and Column. I had the same 
issue when using TIGR Spotfinder to quantitate my images. If you sort 
your info by MR, then MC, and then R, that's the same order as the GAL 
should be (B,R,C). R and C are your Row and Column columns. MR/MC 
defines each block (blocks are counted from the top left towards the 
right, across "metarows"... so MR,MC (1,1) is Block 1, MR,MC (1,2) is 
Block 2, etc... If you write your GAL file starting with the columns: 
Block Column Row Name ID... as in an original GAL file, then that's it.
Your data has to be sorted the same way, of course.

The GAL file only has to be created once. And for the data resorting 
(if necessary) you could just make a little script to re-format the 
input everytime automatically, before you feed it to read.maimages.

Then just read the data specifying the actual column names:

RG<-read.maimages(targets$FileName, columns=list(Gf="F532 Median", 
Gb="B532", Rf="F635 Median", Rb="B635"))

This works for me... I use Genepix and/or SpotFinder data.

Jose

-- 
Dr. Jose I. de las Heras                      Email: J.delasHeras at ed.ac.uk
The Wellcome Trust Centre for Cell Biology    Phone: +44 (0)131 6513374
Institute for Cell & Molecular Biology        Fax:   +44 (0)131 6507360
Swann Building, Mayfield Road
University of Edinburgh
Edinburgh EH9 3JR
UK