[BioC] Mapping Image ID's and Affy probesets

[Robert Gentleman]

Hi Aedin,

Aedin wrote:
> Thanks for the info on my last question. Can I please ask another ;-))
> 
> What do you recommend to use when mapping IMAGE ID's and Affy ID's.
> UniGene and locus link are returning some very strange results (not a
> surprise really :-)).  
> 

  It depends a lot on what you are trying to do. If you are trying to 
combine different data sets, then you probably want to match on sequence 
- which sounds like what you are suggesting below.

  You might also want to look at MergeMaid, and accompanying papers, as 
they have an interesting approach that might help.

  I don' think that there is an example of taking a sequence, say for a 
cDNA and then finding whether a corresponding affy probe set map to a 
similar region, but once you are done I'm sure you will contribute one :-)

  I would use Biostrings, and the corresponding Affy probe package - and 
do exact matching.

   If you end up with many-to-many or many-to-one matches I would match 
them all and then pick the best one.

some code snippets,

  library(annotate)
  library(hgu95av2probe)

  ##here you would use the Acc. Num for the gen of interest
  myseq = getSEQ("D45132")

  ##here you would use the Affy probe id, for the putative match
   wp= hgu95av2probe$Probe.Set.Name == "316_g_at"
  myPr = hgu95av2probe[wp,]

  nchar(myseq)

library(Biostrings)
mybs = NucleotideString(myseq, "DNA")

##do this once for each of the probes - and count how many actually ##match
match1 = matchDNAPattern(as.character(myPr[1,1]), mybs)
as.matrix(match1)

  Does that help?

  Robert

> I know match probes maps between Affy batches.  Can I load IMAGE
> sequences into this to map to a Affy cdf package.  Is there any example
> of this in the BioC help?
> 
> Thanks again BioC for all your help,
> Aedin
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
>