[BioC] Re; MAGEML

Steffen Durinck sdurinck at ebi.ac.uk
Thu Nov 6 10:31:54 MET 2003


Hi Cary,


There should be no problem to import this dataset.  In ArrayExpress 
you'll find most of the time two files that are related to a certain 
experiment.  One contains the experiment data and begins with an 'E' . 
In your case this is E-CAGE-*.  This is also the file you find when 
searching for an experiment via ArrayExpress.  
The second xml file contains all information on the ArrayDesign and 
begins with an 'A'  you'll find this file by clicking on the link 'Array 
design used' which you'll see when you searched for an experiment.
 For you this is A-MEXP-10.

So now you'll have two MAGEML files from which the combined information 
results in a marrayRaw object.

First get the array design info and  make a layout and gnames object 
with the command widget.readMageDesign.  Here you'll have to point to 
both files, see manual. You can use the show packages button to see if 
the package that is required is present in the file you pointed to with 
browse.

Secondly make the raw object by using widget.readMageRaw.
Here you'll only have to use the xml file from  E-CAGE ...

Let us know if you still run into problems.

Cheers,
Steffen & Joke




> Hi all,
>   I am attempting to use public MAGEML data without success.  My current
> datasets are E-CAGE-{2,3,4} from ArrayExpress.  According to the tutorial
> a MAGEML file requires packages
> BioAssayData
> BioAssay
> BioMaterial
> BioSequence
> ArrayDesign
> DesignElement
>
> I suppose this is the problem because inspection of package content shows
> this:
> > magePackageDetector(ecage3)
>  
>  File  aVIB-Col_vs_aUNIL-Col.xml  contains the following packages:
>  
>          BioMaterial_package
>          BioAssay_package
>          QuantitationType_package
>          BioAssayData_package


>
>
> There is only 1 xml file included with the data.  How can I get the
> correct packages?   Any help will be appreciated.  Thanks
>   --cam
>
> -- 
> //=\ Cary Miller                                    //=\ 
> \=// Center for Computational Pharmacology          \=//
> //=\ University of Colorado Health Sciences Center  //=\ 
> \=// http://compbio.uchsc.edu/Hunter_lab/Miller     \=//
> //=\ (303) 315-1584                                 //=\



More information about the Bioconductor mailing list