[Bioperl-l] Bio::Tools::Run::RemoteBlast not getting all hits

Wiepert, Mathieu Wiepert.Mathieu@mayo.edu
Mon, 23 Sep 2002 21:07:42 -0500


HI,

At the very least we need to know some sequences that you are starting with, and the parameters you were setting up with.  If you want to send me the script offline, with sequences, I promise not to post it ;-)  

-Mat

-----Original Message-----
From: Jason Stajich [mailto:jason@cgt.mc.duke.edu]
Sent: Monday, September 23, 2002 8:31 PM
To: Dan Kortschak
Cc: bioperl-l@bioperl.org
Subject: Re: [Bioperl-l] Bio::Tools::Run::RemoteBlast not getting all
hits


Dan - Did you blast these sequences by hand at ncbi and see what you get?
i.e. did anything popping out as strange in those that are repeatably
failing (like 0 hits)?

Mat Wiepert has recently worked with the Tools::Run::RemoteBlast some
more, perhaps we/he/you can try running one of the offending ESTs through
with the latest code and see if things are clearer?

Alternatively you could try running those offending ones through the
netblast (blastcl3) that NCBI provides and post process them with a
script (small problems with bioperl parser and the netblast output but
it is manageable I think).  Hacky, but might get your dataset faster.

-jason

On Tue, 24 Sep 2002, Dan Kortschak wrote:

> Hi All, I have been using Bio::Tools::Run::RemoteBlast to grab sequence
> identities/species/etc for a largish set of coral EST to get an idea of
> sequence conservation in the cnidarian lineage relative to vertebrates and
> non-vertebrate bilaterians (insects and nematodes etc). After running the
> script (largely based on Jason Stajich remote_blast example script - my
> mess of it is pretty ugly so I haven't posted it) and manually validating
> a proportion of the EST blast reports, I have found that a number of EST
> failed to get blast matches and did not report a blast failure, but did
> get matches when blasted manually. These failures were repeatable.
> Does anyone have an idea what may be going on here.
>
> thanks
> Dan
>
> I will post the script if necessary, but it's very ugly and kind of
> embarassing, so I'd rather not.
>
> I'm using bioperl 1.0 (debian wooby release).
>
>

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

_______________________________________________
Bioperl-l mailing list
Bioperl-l@bioperl.org
http://bioperl.org/mailman/listinfo/bioperl-l