[Bioc-devel] Memory problem in affy package of BioC 1.7 and BioC 1.8

[Mirjam van Vroonhoven]

Dear bioconductor developer community,

We have an SGI Origin computation server (32 cpu, 64 Gb RAM) running irix
6.5.27.  We had previously R 2.0.0 with Bioconductor 1.5, and recently we
upgraded to R 2.2.1 with BioConductor 1.7. I tried posting this message to the
bioc-devel list before I subscribed myself but this didn't work. After
suggestion of the list moderator, I also installed R 2.3.0 with BioConductor
1.8 today.

In the newer versions of R we receive memory errors in the affy library, in the
ReadAffy() and the mas5() command, and we also saw them in the vsn
normalization from the vsn library. The same commands run fine in R 2.0.0, on
the same CEL files. It depends on the set of CEL files used, whether we see
errors or not, i.e. enlarging or reducing the set of CEL files with one
or more arrays may introduce the error.

example output in R 2.2.1:
  R : Copyright 2005, The R Foundation for Statistical Computing
  Version 2.2.1  (2005-12-20 r36812)
  ISBN 3-900051-07-0
    <snipped the default R message>
  > library(affy)
  Loading required package: Biobase
  Loading required package: tools
  
  Welcome to Bioconductor
  
          Vignettes contain introductory material.
          To view, simply type 'openVignette()' or start with 'help(Biobase)'.
          For details on reading vignettes, see the openVignette help page.
  
  > data<-ReadAffy()
  > data2<-mas5(data)
  background correction: mas
  PM/MM correction : mas
  expression values: mas
  background correcting...Error: cannot allocate vector of size 211702 Kb
  >


whereas the same commands on the same set of CEL files yield in R 2.0.0:

  R : Copyright 2004, The R Foundation for Statistical Computing
  Version 2.0.0  (2004-10-04), ISBN 3-900051-07-0
    <snipped the default R message>
  > library(affy)
  Loading required package: Biobase
  Loading required package: tools
  Welcome to Bioconductor
           Vignettes contain introductory material.  To view,
           simply type: openVignette()
           For details on reading vignettes, see
           the openVignette help page.
  Loading required package: reposTools
  > data<-ReadAffy()
  > data2<-mas5(data)
  background correction: mas
  PM/MM correction : mas
  expression values: mas
  background correcting...done.
  54675 ids to be processed
  |                    |
  |####################|
  Warning message:
  package 'hgu133plus2cdf' was built under R version 2.2.1
  >

And in R 2.3.0 it yields also an error:
  R : Copyright 2006, The R Foundation for Statistical Computing
  Version 2.3.0 (2006-04-24)
  ISBN 3-900051-07-0
    <snipped the default R message>
  > library(affy)
  Loading required package: Biobase
  Loading required package: tools
  
  Welcome to Bioconductor
  
  
      Vignettes contain introductory material.
  
      To view, simply type 'openVignette()' or start with 'help(Biobase)'.
  
      For details on reading vignettes, see the openVignette help page.
  
  
  Loading required package: affyio
  > data<-ReadAffy()
  > data2<-mas5(data)
  background correction: mas
  PM/MM correction : mas
  expression values: mas
  background correcting...Error: cannot allocate vector of size 188830 Kb
  > 



The memory error occured while 37 Gb of RAM was free.
Memory usage		after 		in mas5() command
			ReadAffy()	
			command
R 2.0.0 / BioC 1.5 	1070M		1189M - 1854M - 2060M
R 2.2.1 / BioC 1.7	 851M		1220M - crash
R 2.3.0 / BioC 1.8	1058M		1242M - crash

I'm puzzled by this....
Has anybody else seen this behaviour?  Is this a bug in the affy package? Is
there a place that I can submit bug reports?

						With kind regards,


						Mirjam van Vroonhoven
-- 
Dr. Mirjam van Vroonhoven
	system administrator/programmer, dept. of Bioinformatics
	Erasmus Medical Center, Rotterdam, The Netherlands
	Room Number Ee 15.32, phone +31-10-463 81 11
Web: 	http://www.erasmusmc.nl/bioinformatics/
E-mail: m.vroonhoven at erasmusmc.nl