[BioC] beadarray: Running BASH for 120 sections

Gavin Koh gavin.koh at gmail.com
Mon Apr 4 22:51:58 CEST 2011


I have 60 samples which were run on an Illumina HumanWG-6 v3.0
Expression BeadChip (so 120 sections) and I am doing the
pre-processing using beadarray.

I am trying to generate spatial masks using BASH(). I have
successfully run a smaller analysis (one slide of 12 sections) on my
MacBook OSX Snow Leopard with 4Gb RAM using beadarray 2.7.

The command I used to call BASH was:
BASHoutput <- BASH(beadData, array=1:n)

I am running the full analysis (120 sections) on a computing cluster
(lustre). I have only requested a single core with 16Gb RAM, because I
don't know how to get BASH() to multithread (although in theory it
ought to be possible? it is a repetitive process after all). I cannot
get the script past 53 sections, without bash() terminating with exit
code "user code 2". Doesn't matter if I am running it interactively or
calling R CMD BATCH. I don't know what the exit code means, so I don't
know how to fix it. I don't think it is out of memory, because lustre
has other codes for reporting out-of-memory and R usually reports
out-of-memory errors as "cannot allocate vector of size..."? Also, the
previous time it ran out of memory (when I tried 6 Gb RAM), it was
lustre that terminated the process.

I don't know if the problem is that BASH() cannot handle so many
sections. If that is in fact the problem, then there are two solutions
I can think of: 1. get BASH() to run multithreaded, or 2. run BASH()
on selected sections only.

On inspection of the pseudoimages, I can see there are only two
sections of the 120 with obvious spatial defects (they look like
scratches). Is it possible to find outliers on the other sections
using the usual (faster) method (>3MAD) and then just use BASH() for
the two sections that are defective only? or...is there a tool to just
draw the masks myself??

Thanks in advance,

Gavin

sessionInfo() reports:
R version 2.12.0 (2010-10-15)
Platform: x86_64-unknown-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=en_GB.UTF-8       LC_NUMERIC=C
 [3] LC_TIME=en_GB.UTF-8        LC_COLLATE=en_GB.UTF-8
 [5] LC_MONETARY=C              LC_MESSAGES=C
 [7] LC_PAPER=en_GB.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] beadarray_2.0.6 Biobase_2.10.0

loaded via a namespace (and not attached):
[1] limma_3.6.6

-- 
Hofstadter's Law: It always takes longer than you expect, even when
you take into account Hofstadter's Law.
—Douglas Hofstadter (in Gödel, Escher, Bach, 1979)



More information about the Bioconductor mailing list