[Biojava-l] BlastLikeSAXParser and Blastall version

Thomas Down td2@sanger.ac.uk
Thu, 2 Jan 2003 12:54:58 +0000


On Tue, Dec 24, 2002 at 12:17:27PM -0800, Rene Warren wrote:
> Hi,
> I have used the BlastLikeSAXParser class and I seem to have set up 
> everything correctly.  However, it catches this exception and I find it 
> quite strange considering that I'm using an older version of blastall 
> (2000).  I also tried a blast output from blastall v 2.2.2 (current) 
> without success.  The same is observed when blastn is used.
> sax exception:org.xml.sax.SAXException: Program ncbi-blastp Version 
> 2.0.14 is not supported by the biojava blast-like parsing framework
> Do you have any thoughts, comments or ideas regarding this?

You can try calling BlastLikeSAXParser.setModeLazy(), to indicate
that the parser should continue even if it doesn't recognize the
precise version of blast which you are using.  My experience of
this is generally good, but obviously you should double-check the
first few results to make sure everything is working.

I note from the documentation of the CVS version of biojava
that 2.2.2 (and 2.2.3) are now listed as supported versions.
If you're using an old version of BioJava which doesn't support
2.2.2, you might want to consider upgrading.

> Also, I have used the SAX XMLparser to parse a blast output in xml 
> format.  Do you guys know if NCBI is complient to the XML format, 
> because that doesn't seem to work either.

You can dump an XML representation of your blast output using
the nativeapps.BlastLike2XML script (in the demos directory).
But this XML is different from NCBI blast XML.  David Huen
has written some BioJava support for the NCBI XML format
(in very recent CVS versions, package org.biojava.bio.program.sax.blastxml).
I don't have direct experience of this myself, though.