[aroma.aff​ymetrix] Error when converting CustomCDF to binary format!

64 views
Skip to first unread message

sean nj

unread,
May 7, 2012, 4:07:36 PM5/7/12
to aroma-af...@googlegroups.com
Hi guys,
 
I tried to convert recent customCDF HTHGU133A_Hs_ENTREZG.cdf (version 15.10) to binary format and end up with error message:
 
Error in convertCdf("HTHGU133A_Hs_ENTREZG.cdf", "HTHGU133A_Hs_ENTREZG,Binary.cdf",  :
  An inconsistency between source and destination CDF was detected. Reason:Units:
Any suggestion?
 
Thanks,
 
Ying
 
> library(affxparser)
> convertCdf("HTHGU133A_Hs_ENTREZG.cdf","HTHGU133A_Hs_ENTREZG,Binary.cdf",
+ version = 4,force = TRUE,verbose=TRUE)
Reading CDF header...
Reading CDF header...done
Reading CDF QC units...
Reading CDF QC units...done
Reading CDF units...
Reading CDF units...done
Writing CDF structure...
Timing for writeCdf():
   user  system elapsed
  10.11    0.02   10.14
Writing CDF structure...done
Comparing CDFs...
Comparing CDFs...
  CDF 1: ./HTHGU133A_Hs_ENTREZG.cdf
  CDF 2: ./HTHGU133A_Hs_ENTREZG,Binary.cdf
  Comparing CDF headers...
  Comparing CDF headers...done
  Comparing QC units...
  Comparing QC units...done
  Comparing units...
Error in convertCdf("HTHGU133A_Hs_ENTREZG.cdf", "HTHGU133A_Hs_ENTREZG,Binary.cdf",  :
  An inconsistency between source and destination CDF was detected. Reason:Units:
> sessionInfo()
R version 2.15.0 (2012-03-30)
Platform: x86_64-pc-mingw32/x64 (64-bit)
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] affxparser_1.28.0   BiocInstaller_1.4.4
loaded via a namespace (and not attached):
[1] tools_2.15.0
>
 

Pierre Neuvial

unread,
May 8, 2012, 3:34:43 AM5/8/12
to aroma-af...@googlegroups.com
Hi Ying,

This is a question for the Bioconductor mailing list, because it had
to do with the affxparser package (and nothing to do with the aroma
framework). Can you post it there ?

Cheers,

Pierre
> --
> 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-af...@googlegroups.com
> To unsubscribe and other options, go to http://www.aroma-project.org/forum/

Henrik Bengtsson

unread,
Jan 7, 2013, 10:37:01 PM1/7/13
to aroma-af...@googlegroups.com, njs...@gmail.com
Just to close this (old) thread:

On Mon, May 7, 2012 at 1:07 PM, sean nj <njs...@gmail.com> wrote:
> Hi guys,
>
> I tried to convert recent customCDF HTHGU133A_Hs_ENTREZG.cdf (version 15.10)
> to binary format and end up with error message:
>
> Error in convertCdf("HTHGU133A_Hs_ENTREZG.cdf",
> "HTHGU133A_Hs_ENTREZG,Binary.cdf", :
> An inconsistency between source and destination CDF was detected.
> Reason:Units:
> Any suggestion?

This happened because:

1. That ASCII CDF (HTHGU133A_Hs_ENTREZG.cdf) from BrainArray
[http://brainarray.mbni.med.umich.edu/Brainarray/Database/CustomCDF/15.1.0/entrezg.asp]
claims there are 12,074 units (NumberOfUnits=12074), but there are
really only 12,073 units. This is related to thread 'convertCdf()
crashes R while converting BrainArray custom CDF for Affy
HT_HG-U133_Plus_PM' on July 3, 2011 [http://goo.gl/QgRsW].

2. Because of the above, convertCdf() will read that false extra unit
as an empty unit that has an 'unknown' unit type.

3. In affxparser (<= 1.30.1), an 'unknown' unit type was incorrectly
encoded by convertCdf() making the unit type appear as a 'copynumber'
unit. This difference is what is detected in the 'Comparing CDFs'
step below.

It's not a serious problem that the unit type is incorrect for this
false extra unit, because it is empty (no groups and no probes)
anyway.

FYI, the encoding error in convertCdf() has been fixed in affxparser
v1.30.2 (just submitted to Bioconductor).

Case closed

/Henrik
Reply all
Reply to author
Forward
0 new messages