[BioC] RMA/QuantileNormalization results difference between oligo and aroma.affymetrix for Hugene

Benilton Carvalho beniltoncarvalho at gmail.com
Fri Feb 26 18:40:19 CET 2010


If you're using the latest oligo, you can use the (not exposed - still
working on further details) function getFidProbeset() :

probeInfo = oligo:::getFidProbeset(rawdata)
idx = probeInfo[["fid"]]
## probeset names are in probeInfo[["fsetid"]]
intensities = exprs(rawdata)[idx,]

and work with 'intensities' (which includes the PMs and controls).

If you rather use only PMs:

pms = pm(rawdata)
pns = probeNames(rawdata)

these are now regular matrices and you can use
rma.background.correct() and normalize.quantiles().

best,

b

On Fri, Feb 26, 2010 at 5:13 PM, Mikhail Pachkov <pachkov at gmail.com> wrote:
> On Fri, Feb 26, 2010 at 5:48 PM, Benilton Carvalho
> <beniltoncarvalho at gmail.com> wrote:
>> Dear Mikhail,
>>
>> when you read the CEL files, you have the information at the probe
>> level. So, when you summarize, you will summarize to a 'less granular'
>> level (in this case probeset).
>>
>> Maybe what you're trying to achieve is the comparison of the
>> background correction (or possibly normalization) method *before*
>> summarization?
>
> You are right we need expression data on *probe-level* without summarization.
>
> Comparison is not my goal. I am just trying to learn how to use these
> packages and choose one for our project. The discrepancy between
> results was very confusing so I asked for help. Obviously the reason
> was my ignorance.
>
> In our project we need to read cel files and get expression values (bg
> corrected and normalized) on *probe-level* for further analysis.
>
> Best regards,
>
> Mikhail
>



More information about the Bioconductor mailing list