Re: [svtoolkit-help] Having trouble with SVPreprocess

[Bob H. <han...@br...>]

Hi, Oliver,

First of all, the "mysample/mysample/null" is OK.
That is actually sample/library/readgroup (and in this case you are 
grouping by library, which is the default and usual thing to do).

Second, there is a known bug, which we are in the process of fixing, 
where some of the preprocessing steps do not catch and report I/O errors
(they use java PrintWriter but do not check for errors).  But since you 
are checking the output file, that is probably not the issue here.
But just fyi.

Third, we have seen occasional problem with this happening with LSF (but 
it is random and occasional and goes away when you rerun).
I don't know if shell mode has the same problem, since I never use that.
One thing to try is --disableJobReport, if you are not already doing 
that.  Also check to see if Queue wrote a .*.done file for this step.
If those things are going wrong, it is a queue problem.

Also, we successfully run LSF 9.1 here with Drmaa.  We had to have our 
IT staff install the LSF Drmaa library.
But it did take a while to get it working.  I think they had to try 
different versions of the drmaa library.
Seva Kashin worked on that and may be able to help if that would be useful.

-Bob

On 2/25/15, 12:02 PM, Homann, Oliver wrote:
>
> Hello,
>
> I'm running into difficulties trying to run SVPreprocess, and was 
> hoping you might have some ideas about what has gone wrong.  After a 
> while the run fails with a (seemingly uninformative) 
> ReduceInsertSizeHistograms error (see end of e-mail). Oddly, I can run 
> that exact same command at the command line (just changing the output 
> name), and it works without reporting any errors.  In both cases an 
> output file of the same size is created, which makes me wonder if this 
> is even the step that is truly failing.
>
> A few other observations/comments:
>
> ·I note that the log appends 'null' to the name of the sample it is 
> processing... is this a clue to something being amiss?
>
> ·I've had to build my own genome bundle because my alignments include 
> the hs37d5 decoy "chromosome".  Perhaps I did something odd there that 
> is contributing to the problem?
>
> ·I am doing this test using the Shell mode since I can't seem to get 
> Queue to work with LSF 9 and drmaa.
>
> Any suggestions would be greatly appreciated.
>
> Regards,
>
> Oliver Homann
>
> ==================================================================
>
> Error from Shell SVPreprocess run
>
> ==================================================================
>
> ERROR 17:34:50,817 FunctionEdge - Error:  'java'  '-Xmx2048m' 
> '-XX:+UseParallelOldGC'  '-XX:ParallelGCThreads=4' 
> '-XX:GCTimeLimit=50'  '-XX:GCHeapFreeLimit=10' 
> '-Djava.io.tmpdir=/mypath/scripts/.queue/tmp'  '-cp' 
> '/myjars/svtoolkit_versions/svtoolkit/lib/SVToolkit.jar:/myjars/svtoolkit_versions/svtoolkit/lib/gatk/GenomeAnalysisTK.jar:/myjars/svtoolkit_versions/svtoolkit/lib/gatk/Queue.jar' 
> '-cp' 
> '/myjars/svtoolkit_versions/svtoolkit/lib/SVToolkit.jar:/myjars/svtoolkit_versions/svtoolkit/lib/gatk/GenomeAnalysisTK.jar:/myjars/svtoolkit_versions/svtoolkit/lib/gatk/Queue.jar' 
> 'org.broadinstitute.sv.apps.ReduceInsertSizeHistograms' '-I' 
> '/mypath/output/isd/mysample.hist.bin'  '-O' 
> '/mypath/output/isd/mysample.dist.bin'
>
> ERROR 17:34:50,830 FunctionEdge - Contents of 
> /mypath/output/logs/SVPreprocess-6.out:
>
> INFO 16:35:09,560 HelpFormatter - 
> -------------------------------------------------------------------
>
> INFO 16:35:09,562 HelpFormatter - Program Name: 
> org.broadinstitute.sv.apps.ReduceInsertSizeHistograms
>
> INFO 16:35:09,566 HelpFormatter - Program Args: -I 
> /mypath/output/isd/mysample.hist.bin -O 
> /mypath/output/isd/mysample.dist.bin
>
> INFO 16:35:09,569 HelpFormatter - Executing as me...@my... on 
> Linux 2.6.32-358.el6.x86_64 amd64; Java HotSpot(TM) 64-Bit Server VM 
> 1.7.0_25-b15.
>
> INFO 16:35:09,570 HelpFormatter - Date/Time: 2015/02/24 16:35:09
>
> INFO 16:35:09,570 HelpFormatter - 
> -------------------------------------------------------------------
>
> INFO 16:35:09,570 HelpFormatter - 
> -------------------------------------------------------------------
>
> Processing mysample/mysample/null ...
>
> INFO 16:35:20,573 CommandLineProgram - Program completed.
>
> ==================================================================
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the
> conversation now. http://goparallel.sourceforge.net/
>
>
> _______________________________________________
> svtoolkit-help mailing list
> svt...@li...
> https://lists.sourceforge.net/lists/listinfo/svtoolkit-help