[Bioperl-l] organism restricted blast search

[Jason Stajich]

I believe there are ways of building virtual subsets of the databases
locally with some tricks with formatdb but I don't really know them.

You can also parse your reports and lookup the GI number using the tools
in Bio::Taxonomy and Bio::DB::Taxonomy to get the species name and or
restrict hits to a particular subset of the taxonomy hierarchy.  There
were some posts in the last few months about this.  You can re-write your
report with Bio::SearchIO::Writer::TextResultWriter or HTMLResultWriter
and feed in a function which will only output a hit if it falls in a
particular domain of the taxonomy. That is a lot of shorthand as to how
program it up - if you really need something like this can point you to
more working code.


-jason
On Thu, 5 Aug 2004, Fernan Aguero wrote:

> +----[ Pedro Antonio Reche <reche at research.dfci.harvard.edu> (05.Aug.2004 10:28):
> |
> | Hi,
> | I am interested in carrying out a blast (or parse a blast search) in  a
> | way that only hits from a given organism will show. This is similar to
> | what the NCBI blast site does currently I wonder if someone knows how
> | to do it using bioperl. Thanks in advance for any help.
> | Cheers
> |
> +----]
>
> If you use Bio::Tools:Run::RemoteBlast to run your searches
> remotely at the NCBI, you can do something along the lines of:
>
> #change a parameter
> $Bio::Tools::Run::RemoteBlast::HEADER{'ENTREZ_QUERY'} = 'Homo sapiens [ORGN]';
>
> this was taken from the synopsis of the mentioned module.
> You can of course replace your entrez query with whatever
> you like.
>
> If you're doing local blast searches, I guess it's not
> possible. If anyone has a way of doing it, please enlighten
> me.
>
> Fernan
>
>

--
Jason Stajich
Duke University
jason at cgt.mc.duke.edu