Blast Text-Output To Xml Converter

Question

[Leszek]

Hi, I'm looking for simple way of converting blast output in text (stardard output) into XML format (maybe some has got script?). I've looked elsewhere but couldn't find anything:/

UPDATE
I ended up running blastall with -m7, but stil someone may benefit from BlastConverter :)

[Egon Willighagen]

Please add some example output... Why do you want it as XML anyway?

[Lythimus]

I would be interested in a solution to this as well. My BLAST+ run is eating all 64GB of my RAM in BLASTXML format. I have to have BLASTXML format to support MEGAN, but if I can convert a standard run to BLASTXML that would be ideal--assuming it stores data less verbosely in RAM.

Answer 1

[Pierre Lindenbaum]

did you know that the 'blastall' command line has an option (-m 7) for generating an XML output ? http://www.ncbi.nlm.nih.gov/staff/tao/URLAPI/blastall/blastall_node29.html

If you're using the web interface of the NCBI for blast, the result can be downloaded as XML (or ASN1,... )

most Bio* packages can parse blast (e.g. http://www.bioperl.org/Core/Latest/bptutorial.html#iii_4_2_parsing_blast_and_fasta_reports_with_search_and_searchio ) but I don't know if there is a predefined method to export the data as XML.

[Leszek]

I'm aware of that, however I would love to use blast-machine for computation (DeCypher by Timelogic) and it doesn't give xml output, only ncbi like text output... So this is why I'm asking...

Answer 2

[Darked89]

blastall already has an option of producing XML:

blastall -m 7

Simply rerun your blast search instead of writing a parser.

[Bach]

I have the feeling the submitter might not have the possibility to (re-)run that BLAST.

[Leszek]

I would love to use blastall, however running regular blast will take ages... I need to do it on blast-machine and feed xml-formatted output to another step:/

Answer 3

[Dwood]

I to have this problem. Which cannot be solved by re-running blast. I am currently writing a script to do it as I cannot find any solutions online. It's slightly annoying that the default text/html output seems to be the least amenable to conversion. If I come up with anything useful I'll make sure it is made available.

Answer 4

[Anon1234Y]

Extremely simple task:

<xml>
  <blast>
  <![CDATA[
    copy your blast result here.
  ]]>
  </blast>
</xml>

Easy way for valid xml when no XML Schema is specified :P

[Leszek]

this is not going to solve the problem at all

Answer 5

[Hamish]

For a script to convert the default NCBI BLAST text output (as specified with '-outfmt 0' for NCBI BLAST+ or '-m 0' for legacy NCBI BLAST) into NCBI BLAST XML, see the following thread in the Blast2GO Google Group:

https://groups.google.com/forum/#!topic/blast2go/zJJCN-timCg