[BioC] filtered CDF file results in errors

[lgautier at altern.org]

> Sebastien Gerega wrote:
>> Hi,
>> I am working on the yeast2 affymetrix chip and am trying to filter out
>> all the s_pombe probes with the following code:
>>
>> library(altcdfenvs)
>> bothcdf = wrapCdfEnvAffy(yeast2cdf, 496, 496, "yeast2cdf")
>> pp = read.table("s_pombe.msk", sep="\t", skip=2)
>> ids = as.character(pp[,1])
>> saccdf = bothcdf[ids]
>> print(bothcdf)
>> print(saccdf)
>> saccdf <at> envName = "Sac ids only"
>> envsaccdf = as(saccdf, "environment")
>> data = ReadAffy(filenames=filenames)
>> data at cdfName <- "envsaccdf"
>>
>> "s_pombe.msk" was downloaded from the Affymetrix website and contains
>> a list of all the s_cerevisiae probes (those I want to use).
>>
>> This code runs without problems but once I get to QC I get the
>> following error:
>> hist(data, main="Log2(Intensity)", col=targetCol)
>> Error in density.default(newX[, i], ...) : 'x' contains missing values
>>
>> I have tried running the reverse analysis, that is filtering out the
>> s_cerevisiae probes instead of the s_pombe ones and I no longer get
>> the error.
>> What could be going wrong? Can anyone give me some tips to
>> troubleshoot the problem?
>> thanks in advance,
>> Seb
>>
>>
>> I have tried again using the new BioC release and no longer get an
>> error when creating the histogram.
>> However, now when I run fitPLM on the AffyBatch object R completely
>> crashes. If I run fitPLM on the AffyBatch without having changed the
>> CDF environment I get no problems.
>> Does anyone have any ideas what might be causing this? or how I could
>> deal with the problem?
>> thanks,
>> Seb
>>
>>
> I seem to have resolved this problem by removing the probes listed in
> "s_cerevisiae.msk" rather than using those listed in "s_pombe.msk" with
> the following code:
>
> bothcdf = wrapCdfEnvAffy(yeast2cdf, 496, 496, "yeast2cdf")
> pp = read.table("s_cerevisiae.msk", sep="\t", skip=2)
> ids = geneNames(bothcdf)[which(geneNames(bothcdf) %in% pp[,1] == FALSE)]
> saccdf = bothcdf[ids]
> print(bothcdf)
> print(saccdf)
> saccdf at envName = "Sac ids only"
> envsaccdf = as(saccdf, "environment")
>
> However, I am still curious what was causing the problem....
>


Could you make available the files needed to reproduce the problem ?
(we can sort the details of data transfer off-list).


Thanks,


L.