[Biopython-dev] RPS-BLAST XML output - Was: Biopython new release coming up

[Peter]

Peter wrote:
> In BioPython 1.40b there where no tests of NCBIXML.py
> 
> ...

Thanks due to Michiel for adding these to BioPython with a test script.

> Once this is pinned down, I want to try and try dealing with some 
> RPS-BLAST XML output with NCBIXML.py...

See bug 1715 for parsing the rpsblast txt output:

http://bugzilla.open-bio.org/show_bug.cgi?id=1715

(Summary - I had a cobbled together a working text parser for RPS-BLAST 
2.2.10, but it would need changing for RPS-BLAST 2.2.12)

Attached to this email are two simple examples using standalone 
RPS-BLAST 2.2.10 and 2.2.12, with both txt and XML output.

'xbt006.xml', # Standalone RPS-BLAST 2.2.10, 
gi|49176427|ref|NP_418280.3| - PFAM database
'xbt007.xml', # Standalone RPS-BLAST 2.2.12, 
gi|49176427|ref|NP_418280.3| - PFAM database
'xbt008.xml', # Standalone RPS-BLAST 2.2.10, 
gi|729325|sp|P39483|DHG2_BACME - CDD database

The NCBIXML.py seems to work fine with the standalone RPSBLAST XML 
output.  Note that the online RPS-BLAST does not seem to offer XML 
output at the moment...

One very odd thing I noticed, is that the XML files from RPS-BLAST seem 
to claim to have been produced by blastp:

   <BlastOutput_program>blastp</BlastOutput_program>
   <BlastOutput_version>blastp 2.2.10 [Oct-19-2004]</BlastOutput_version>

Or:

   <BlastOutput_program>blastp</BlastOutput_program>
   <BlastOutput_version>blastp 2.2.12 [Aug-07-2005]</BlastOutput_version>

The text output correctly states RPS-BLAST, so this looks like an 
RPS-BLAST bug which I plan to report...

Peter
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xml_rps_blast_test_files.zip
Type: application/x-zip-compressed
Size: 17889 bytes
Desc: not available
Url : http://portal.open-bio.org/pipermail/biopython-dev/attachments/20051028/ddc47ee3/xml_rps_blast_test_files-0001.bin