[BioC] cdf-package to .cdf-file using affxparser

[Kasper Daniel Hansen]

In that case I would obtain the original CDF file from Affymetrix. You  
can use the header of that file as a template. You will also get all  
the extra information of all the probes. You will see that the output  
is essentially a list of lists of lists... and all the re-annotation  
does is essentially re-ordering these lists. That should be easy to  
figure out. In addition you can use the QC probesets from the original  
CDF files, if you want to include them. One thing I would check is  
whether the same probe occurs in multiple probesets - many routines  
depends on that not being the case - although nothing in the CDF  
specification requires that a probe only belongs to one probeset.

Kasper

On Oct 27, 2008, at 3:31 , Samuel Wuest wrote:

> Dear Kasper,
>
> thanks for your help. I understand that most cdf-packages are  
> packaged Affy-.cdf files; however, I am using the "tinesath1cdf"  
> package - downloaded from the Bioconductor homepage, see "experiment  
> data" directory-  which is a custom-made reannotation of the Ath1- 
> Arabidopsis GeneChip (and according to the author, no .cdf-file is  
> available). Therefore I was wondering whether there is a direct way  
> from R-package to .cdf-file.
>
> You have suggested to read-in the affymetrix .cdf-file, now I have  
> done this already and found that there are some identifiers which  
> are not available in the custom-made package. However, the probes  
> are mapped all right (I have X, Y, sequence and probe-set ID) and  
> the identifiers I am missing are basically arbitrary numbers (e.g.  
> Unit-ids), which I could theoretically just make up myself: if you  
> do not see any obvious problems there...
>
> Anyway, I ll stick to your suggestion to use the readCdf-output as a  
> guidline, thanks a million and best wishes,
> Sam
>
> 2008/10/27 Kasper Daniel Hansen <khansen at stat.berkeley.edu>
> Hi Sam
>
> The easiest way forward would be for you to download a CDF file from  
> affymetrix site and use readCdf on that file and then study the  
> output of that function. The input structure to writeCdf can be  
> complicated at first, but corresponds exactly to the output of  
> readCdf.
>
> Not all information in a CDF file is contained in a CDF package, so  
> you cannot always go from package -> file (although often the  
> additional information in a CDF file is not used - but that is very  
> context dependent).
>
> Why are you interested in doing this? Typically the CDF packages  
> from Bioconductor are just packaged versions of the Affy files, so  
> why not just use them directly.
>
> Kasper
>
>
>
> On Oct 25, 2008, at 8:09 , Samuel Wuest wrote:
>
> Dear List,
>
> I am trying to generate a .cdf-file (that could be used for programs  
> such as
> dChip or aroma.affymetrix) from a cdf-package downloaded from the
> Bioconductor homepage (which provides me an environment containing  
> the probe
> mappings)...
> Is there a direct way to do this (let's say something like
> "writeCdf("myChipcdf-package"))? If not, how can I generate a  
> cdfheader and
> a cdf-list used for the function "writeCdf" from the mappings in the
> environment? (I got hands on a script that uses a flat-file as input  
> for the
> function, but I am having problems generating this flatfile from the R
> environment).
>
> Thanks a million for any help on this.
>
> Best wishes,
>
> Sam
>
> ------------------------------------------------------
> Wuest Samuel
> Smurfit Institute of Genetics
> Trinity College Dublin
> Dublin 2, Ireland <http://www.tcd.ie/Genetics/wellmer-2/index.html>
> ------------------------------------------------------
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>
>
>