[BioC] cdfenv files ATTN: James McDonald

[James W. MacDonald]

Andrew Van Brunt wrote:
> Hello -
> 
> I am almost embarrassed to be posting to this list since my R programming
> skills are essentially non-existent and my computer skills in general a
> probably quite lacking compared to others on this list. But I have been
> banging my head on the wall for too long now so I figured I better suck it
> up and post.

Insert obligatory flame here ;-D

> 
> I apologize in advance for the length of this message but as I said, I am
> a very naive user and I have lots of questions and I also want to present
> some evidence that I've been trying my best to find the answers before
> posting to the list. And the subject does contain James McDonald but thats
> because he seems to reply to questions of this nature and thus seems to be
> something of an expert on this topic but really anyone's help would be
> appreciated.
> 
> The problem I have is we are attempting to use R/BioC to analyze some
> affymetrix SNP chip data. Specifically, we have .cel files and the chip
> used was the Human Mapping 250K_Sty array. Basically, I need to be able to
> associate the probe intensity values with the proper probes. I know this
> can be done with Affy software but we have hundreds of chips and you can
> not export probe intensity tables from programs like GTYPE except 10 at a
> time, so that would be very labor intensive. In the future we may have
> many hundreds or even thousands of chips to analyze and point and click
> will not be an acceptable solution.
> 
> So basically my question is this:
> I need a package that will associate the correct intensities with the
> probe names for further analysis. It doesn't seem that there are really
> any packages for doing this with any mapping arrays as most array analyses
> through BioC (it seems to me anyways) are concerned with expression
> arrays. Are there packages for Mapping arrays?
> here is what I have learned/tried so far:
> 
> visiting this thread ->
> http://article.gmane.org/gmane.science.biology.informatics.conductor/4195/match=chip+libraries
> 
> I learned that not all cdfenv files are listed for download on the BioC
> site. So I tried the recommended solution here using reposTools, only my
> menu does not look like the one posted and there is not an option for the
> CDF Environment Repository. I think this was a pretty old recommendation.
> So does anyone know how I can see those that are not on the BioC website?
> 
> Have also read several threads and the documentation for makecdfenv and I
> am unfortunately on R for windows and I think I have the windows binary as
> opposed to the source code installed. I've read about the extra tools and
> Perl you have to download and all this other stuff and its all very
> confusing to me - needless to say I've tried to make my own library but
> have had no such luck. I even have problems installing the Active perl for
> windows in the right location etc. I've tried using cygwin and not had
> much luck there either. So what I really need is a zipped windows binary
> package that I can install for the Mapping_250K_Sty chip so I can read
> those cel files and work with the probe intensities. So if anyone knows
> where I can find one please let me know. Or if anyone would be willing to
> make one for me they would have my eternal gratitude. I guess lastly, if
> anyone would be willing to walk a real computer novice on how to make the
> environment and create a windows binary package out of it (so I can port
> it to other windows machines) that would be appreciated too... not as much
> perhaps and just giving me a package to download but appreciated
> nonetheless.
> 
> Any takers?
> Any ideas?
> Any suggestions?

It depends on what you really want to do. If you want to actually 
analyze the data from within R, I think you are stuck at this point 
because you are on windows and most likely won't have (or be able to 
use) the amount of RAM required to analyze these things. The correct 
package for analyzing SNP chips is the oligo package, but currently this 
package holds all data in memory, so you will have trouble analyzing 
tens of chips with a 64 bit operating system and say, 16 Gb of RAM, not 
to mention the ~3 Gb of RAM that you can address using a windows box.

If you just want to associate probes with probesets and then analyze 
using some other software, then you might be able to use affxparser. 
This package, like oligo, is still in the development stage, so tends to 
be less documented than a release package. Unfortunately, this means you 
will have to climb R's steep learning curve a bit quickly if you want to 
get anything done. Luckily for you, the developers of both oligo and 
affxparser are quite friendly and tend to be very helpful for those who 
are trying to use their packages.

One other alternative is to use Affy's software. These days you can 
download Affy software for free, and it tends to be directed towards 
those on Windows, and IIRC, can handle lots of SNP chips.

HTH,

Jim


> 
> Thanks,
> Andrew Van Brunt
> Division of Statistical Genomics - Washington University in St. Louis
> 
> reply to: andrew at dsgmail.wustl.edu
> 
> _______________________________________________
> 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


-- 
James W. MacDonald, M.S.
Biostatistician
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623


**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues.