On Wed, Jul 30, 2008 at 5:00 PM, Hailong <Hailo...@asu.edu> wrote:
>
> Hello,
>
> Sorry for the mass email.
No worries, it's 2008 and mail filters are well developed.
>
> What I am doing now is to test several normalization methods (rma,
> dchip, etc) on small data sets (2 CEL files for example) on my MacBook
> Pro. If it works out, I plan to move onto a Linux workstation.
I have to say that doing that on 2 arrays only will give you nothing
but "rubbish" because of the way the models and algorithms are
designed. You should at least work with 5-10 arrays. FYI,
aroma.affymetrix will work the same regardless of platform. Memory
will *not* be an issue.
>
> Installation of the aroma.affymetrix package on mac and linux worked
> fine, as well as the loading of the library.
>
> I am stuck at http://groups.google.com/group/aroma-affymetrix/web/setup.
> So the problem is that I need to download a .CDF file for hgu133plus2,
> and I have not found out how to get a .CDF file for Human Genome U133
> 2.0 Arrays.
From http://groups.google.com/group/aroma-affymetrix/web/frequently-asked-questions:
FAQ. 2007-05-24: Where can I download CDF files?
A: The CDF file for a given chip type is available in so called
"Library Files" at the corresponding "Support Materials" page on the
Affymetrix website. See also Section "Documentation on chip types" on
this site."
Note that you will have to download a big archive file and inside you
will find the *.CDF file among other files.
>
> First I tried install the package at
> http://www.bioconductor.org/packages/2.2/data/annotation/html/hgu133plus2cdf.html.
> I also download the OS X binary hgu133plus2cdf_2.2.0.tgz and
> decompressed it. There was only one file hgu133plus2cdf under /
> hgu133plus2cdf/data, and I don't think this is what I am looking for.
Nope. The Bioconductor packages does not contain CDF files. It is
very unfortunately that even the Bioconductor developers are sloppy
using the term "CDF" when they really mean "CDF environment package".
Note that CDF is an acronym for "Chip Definition File". Here we write
"CDF file" because we want to make it more explicit that it is a file.
Thus, a CDF is exactly one file ending with a file name extension
*.CDF (or *.cdf) and it follows a well defined file format specified
by Affymetrix - nothing else.
>
> After reading previous discussion articles, I also checked the
> following web sites:
>
> http://brainarray.mbni.med.umich.edu/Brainarray/Database/CustomCDF/11.0.0/entrezg.asp
> http://www.affymetrix.com/support/technical/byproduct.affx?product=hg-u133-plus
>
> I know it's a stupid questions, but is there anyone who can give me
> more information on downloading a .CDF file as well as the difference
> between installing a package for CDF and the corresponding .CDF file?
Hope the above helps. Let us know if you still have problems.
Cheers
Henrik
>
> Thank you very much.
> >
>
(Just to follow up on Henrik's post)
I am in the process of making available the CDF files that I have used
in my work with aroma.affymetrix. This is also with the hope that
if people use custom annotation (or use chips that we have not used
yet), that they would also contribute their custom CDFs (or binary-
converted versions of the text CDFs that you can download from Affy).
I have been updating some of the Google group pages for the newer
expression chips and will send a message around to the group in a day
or two about those once they are finished ... so I may as well update
HG-U133 too ...
So .... in the past, I have used the U133 Plus2 chips within
aroma.affymetrix in a couple of ways (using different sources of
annotation e.g. Ensembl, RefSeq or just Affy's). I've updated the
documentation and made available some binary CDFs for this chip. See:
http://groups.google.com/group/aroma-affymetrix/web/hg-u133-plus-2
Hope that helps.
Mark
------------------------------
Mark Robinson
Epigenetics Laboratory, Garvan
Bioinformatics Division, WEHI
e: m.rob...@garvan.org.au
e: mrob...@wehi.edu.au
p: +61 (0)3 9345 2628
f: +61 (0)3 9347 0852
------------------------------