[Biopython] Problems parsing with PSIBlastParser

[Michiel de Hoon]

Last time I checked (which was a few weeks ago), a multiple-query PSIBlast search gives a file consisting of concatenated XML files. The problem is in the design of Blast XML output. For a single-query PSIBlast, the fields under <BlastOutput_iterations> are used to store the output of the PSIBlast iterations. For multiple-query regular Blast, the same fields are used to store the search results of each query. With multiple-query PSIBlast, there is then no way to store the output in the current XML format.
I've been meaning to write to NCBI about this, but I haven't gotten round to it yet. Will do so this weekend.

--Michiel.

--- On Thu, 10/15/09, Miguel Ortiz Lombardia <ibdeno at gmail.com> wrote:

> From: Miguel Ortiz Lombardia <ibdeno at gmail.com>
> Subject: Re: [Biopython] Problems parsing with PSIBlastParser
> To: "Peter" <biopython at maubp.freeserve.co.uk>
> Cc: "Biopython Mailing List" <biopython at lists.open-bio.org>
> Date: Thursday, October 15, 2009, 12:52 PM
> 
> Le 15 oct. 09 à 18:32, Peter a écrit :
> > 
> >> I still believe that the logic behind the
> NCBIStandalone.PSIBlastparser is
> >> correct or, at least, useful. But I could change
> my mind if you think
> >> otherwise.
> > 
> > The idea of the NCBIStandalone.PSIBlastparser plain
> text parser, and
> > its object structure makes sense.
> > 
> 
> Good!
> 
> >> I have filed the bug:
> >> http://bugzilla.open-bio.org/show_bug.cgi?id=2929
> >> and have upload the XML from blastpgp v. 2.2.22
> mentioned above.
> > 
> > Lovely - thank you. So that is a single query, with 3
> iterations.
> > What would be *really* nice, is a multiple query file
> (say three
> > queries, each needing just a few iterations to keep
> the file small).
> 
> 
> Never used multiple query file... Do you mean starting from
> a multiple-alignment file with the -B option?
> 
> -- Miguel
> 
> 
> 
> 
> _______________________________________________
> Biopython mailing list  -  Biopython at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biopython
>