[BioC] Tools for comparing microarray data on a snp level

[Sean Davis]

On Thu, Mar 18, 2010 at 11:05 AM, Vincent Davis
<vincent at vincentdavis.net> wrote:
> First, I am very new to this and have been learning a lot in a 3 week crash
> course involvement a project. My point is that I am not great with the
> terminology.
>
> What I have, ~50 affycel files, and 3 groups of ~8 samples each for which I
> know the genome for, ie I know what probes/sequences should bind. We do not
> have a CDF file as this is a custom array with a mix of probe purposes.
> Basically these 3 groups of 8 are my means of validating the method used
>
> The other ~25 are unknown and I would like a to put a probability on each
> probe/sequence of the likelihood that a sequence that matches. ie is in the
> genome. of the unknown.
>
> I also have single snp mismatch data for each of the known and the location
> of the mismatch in the sequence.
>
>
> Are there any tools in bioconductor and or methods that I should look at. Up
> to this point I have been using python to merge the different data and do
> some initial analysis.

Hi, Vince.  You might have a look at the affy and oligo packages for
data import and normalization possibilities.  I have to admit that I'm
not sure what workflow will be appropriate without a CDF file, though.
 The limma package (and several others) is good for looking at
differential hybridization signal.  R is a fantastic tool for data
exploration, so looking at the effect of snp position and
presence/absence in your data should be possible with basic plotting
and hypothesis tests.

Others with more experience in the field of comparative genomics might
have more detailed thoughts.

Sean