[Biojava-l] Any memory-optimised blast parser available ?

[Andy Yates]

Hi,

Have you tried increasing the amount of memory the Java virtual machine 
has available? The problem with Java is that it starts with a default 
memory limit of approx. 128MB which if you've got a memory intensive 
process will not be enough.

When running your program try adding the option

-Xmx256M

So if my class was called Main.class my command line would be:

java -Xmx256M -cp .:biojava.jar Main

Or if you're running this from ant then there is an attribute maxmemory 
which can be specified in the <java/> task.

Hope that helps

Andy

Benoit VARVENNE wrote:
> Hello,
> 
> I'm trying to parse blast result files from a remote blast on NCBI
> databases.
> In order to do so, i'm using biojava XML parser like this :
> 
> -----------
>       InputStream is = new FileInputStream(blastFile);
>       BlastXMLParserFacade parser = new BlastXMLParserFacade();
>       SeqSimilarityAdapter adapter = new SeqSimilarityAdapter();
> 
>       parser.setContentHandler(adapter);
> 
>       List results = new ArrayList();
> 
>       SearchContentHandler builder = new BlastLikeSearchBuilder(results,
>           new DummySequenceDB("queries"), new
> DummySequenceDBInstallation());
> 
>       adapter.setSearchContentHandler(builder);
> 
> // Here is the method that throws 'OutOfMemoryException'
>       parser.parse(new InputSource(is));
> -----------
> 
> While blasting a genomic sequence (i've tried with 50Kb long sequences max),
> I've got an OutOfMemory exception.
> 
> Does anyone know a way to resolve this ? (this aim is not to dividde my
> genomic locus in too many parts, 50Kb is a quarter of my genomic...)
> 
> Thanks,
> Regards,
> 
> Benoit
> 
>