Hi.
On Fri, Feb 4, 2011 at 6:35 PM, Mete Civelek <mete.cive...@gmail.com> wrote:
> Hi All,
>
> I am following the code in the "Lab: CRMA v2 on 10K tumor data"
> document to process 147 SNP6.0 arrays. 39 of these arrays are from
> females and the rest are males. I am getting the following error
> message at the end of circular binary segmentation step:
>
> Exception: Something is wrong with the data. Too many non-finite
> values: 36890 (25.2% > 20.0%)
> at throw(Exception(...))
> at throw.default(sprintf("Something is wrong with the data. Too many
> non-finite values: %d (%.1f%% >
> at throw(sprintf("Something is wrong with the data. Too many non-
> finite values: %d (%.1f%% > %.1f%%)"
> at getRawCnData.CopyNumberChromosomalModel(this, ceList = ceList,
> refList = rfList, chromosome = chro
> at getRawCnData(this, ceList = ceList, refList = rfList, chromosome
> = chromosome, ..., estimateSd = F
> at extractRawCopyNumbers.CopyNumberChromosomalModel(this, array =
> array, chromosome = chr)
> at extractRawCopyNumbers(this, array = array, chromosome = chr)
> at fit.CopyNumberSegmentationModel(this, ..., .retResults = FALSE,
> verbose = verbose)
> at fit(this, ..., .retResults = FALSE, verbose = verbose)
> at plot.CopyNumberSegmentationModel(model, path = path, imageFormat
> = "png", plotband = plotband, arr
> at plot(model, p
>
> The sessionInfo is below:
>
>> sessionInfo()
> R version 2.11.1 (2010-05-31)
> x86_64-pc-mingw32
>
> locale:
> [1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United
> States.1252
> [3] LC_MONETARY=English_United States.1252
> LC_NUMERIC=C
> [5] LC_TIME=English_United States.1252
>
> attached base packages:
> [1] stats graphics grDevices utils datasets methods
> base
>
> other attached packages:
> [1] RColorBrewer_1.0-2 DNAcopy_1.22.1
> sfit_0.1.9 aroma.affymetrix_1.7.0 aroma.apd_0.1.7
> [6] affxparser_1.20.0 R.huge_0.2.0
> aroma.core_1.7.0 aroma.light_1.16.1 matrixStats_0.2.2
> [11] R.rsp_0.4.0 R.cache_0.3.0
> R.filesets_0.9.0 digest_0.4.2 R.utils_1.5.3
> [16] R.oo_1.7.4 R.methodsS3_1.2.1
>
> loaded via a namespace (and not attached):
> [1] splines_2.11.1
>
> I am assuming that this error is to the presence of two genders but of
> course I could be wrong. Can anyone provide an advice as to how to do
> this properly to avoid getting the same error?
The error is a "sanity check" for the purpose of trying to detect
unexpected results. It may be that the sanity check is to
conservative with your data set. However, it's been long time since I
heard having a problem with the current 20% threshold.
Typically, you would expect to see this problem on Chr X or Chr Y (due
to gender differences) and that was indeed what this test was
initially designed for, because there is where you are more likely to
get estimated CNs that are zero and when taking the log that become a
"non-finite signal". However, from the error message "36890
(25.2%..." one can tell it is a problem detected on a chromosome with
approx 36890/0.252 = 146,389 loci, which match the Chr01 for the GWS6
chip type. So, this indicates that this is an extraordinarily
problem.
Thus, it might very well been an issue with your data, or more
precisely that the preprocessing failed for one or more of your
samples. Note that the error is per sample specific, so it may be
that it is only one sample causing the error.
Let's troubleshoot. Try the following:
cbs <- CbsModel(cesNTotal);
cn <- extractRawCopyNumbers(cbs, array=1, chromosome=1, verbose=-5);
Change 'array' to 2, 3, ... until you find a sample that give an
error. If my above guess on chromosome, you might have to play with
that too.
When you've identified one problematic sample and chromosome, relax the test by:
cn <- extractRawCopyNumbers(cbs, array=1, chromosome=1,
maxNAFraction=0.5, verbose=-5);
The plot the data as:
plot(cn, ylim=c(-6,6));
What does it look like? Is it all over the place? Please send a PNG of that.
Try some other samples. Do you have problems with others too, or just
one/a few?
Let's start with the above.
> Should I run the
> "Vignette: Sex-chromosome bias-corrected reference signals from pooled
> average?"
That is unrelated and will only affect the vertical shift of your CN
estimates on Chr X and Chr Y. It would make a difference on the other
chromosomes.
/Henrik
>
> Thank you very much for your help...
>
> Mete
>
> --
> When reporting problems on aroma.affymetrix, make sure 1) to run the latest
> version of the package, 2) to report the output of sessionInfo() and
> traceback(), and 3) to post a complete code example.
>
>
> You received this message because you are subscribed to the Google Groups
> "aroma.affymetrix" group with website http://www.aroma-project.org/.
> To post to this group, send email to aroma-affymetrix@googlegroups.com
> To unsubscribe and other options, go to http://www.aroma-project.org/forum/
>
--
When reporting problems on aroma.affymetrix, make sure 1) to run the latest
version of the package, 2) to report the output of sessionInfo() and
traceback(), and 3) to post a complete code example.
You received this message because you are subscribed to the Google Groups
"aroma.affymetrix" group with website http://www.aroma-project.org/.
To post to this group, send email to aroma-affymetrix@googlegroups.com
To unsubscribe and other options, go to http://www.aroma-project.org/forum/
