[BioPerl] Re: [Bioperl-l] here's a cool bug!

[Mark Wilkinson]

Hey Jason, 

Actually, let's hold off on making any adjustments.  This genbank file
came from a SeqHound call, and I'm a bit suspicious of it.  I compared
it to the actual genbank record and they are non-identical, though
SeqHound is supposed to (from what I have been told) be providing bona
fide original genbank records, not synthesized ones... so I don't know
why they are different, but the differences are pretty substantial.

I've contacted the SeqHound developers with this question to try to find
out what's up.  If this is *not* a genuine genbank record, then we
probably shouldn't change the code.

M


On Thu, 2003-12-11 at 20:01, Jason Stajich wrote:
> [jason at ascona jason]$ perl -mBio::SeqIO -e 'print
> $sio=Bio::SeqIO->new(-file => "blah.gb")->next_seq->primary_id,"\n";'
> 
>  Bio::Seq::RichSeq=HASH(0x8518394)
> 
> So there is no VERSION for the sequence...
> VERSION       GI:230940
> 
> Is that valid Genbank...
> 
> Need to make this loop here a little tighter then I guess
> in genbank.pm
> 
> 	  #Version number
> 	  elsif( /^VERSION\s+(.+)$/ ) {
> 	      my ($acc,$gi) = split(' ',$1);
> 	      if($acc =~ /^\w+\.(\d+)/) {
> 		  $params{'-version'} = $1;
> 		  $params{'-seq_version'} = $1;
> 	      }
> 	      if($gi && (index($gi,"GI:") == 0)) {
> 		  $params{'-primary_id'} = substr($gi,3);
> 	      }
> 	  }
> 
> Any takers...?  Make that regexp a little more greedy for GI: perhaps?
> 
> -jason
> On Thu, 11 Dec 2003, Mark Wilkinson wrote:
> 
> > This bug is an interesting one!  It doesn't happen in all cases, only on
> > certain genbank records, like the one attached.  Try this out:
> >
> > perl -MBio::SeqIO -e '$sio=Bio::SeqIO->new(-file => "blah.gb");
> > $s=$sio->next_seq; print $s->primary_id;'
> >
> > that should lead to some raised eyebrows :-)
> >
> > M
> >
> >
> >
> 
> --
> Jason Stajich
> Duke University
> jason at cgt.mc.duke.edu
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at portal.open-bio.org
> http://portal.open-bio.org/mailman/listinfo/bioperl-l
-- 
Mark Wilkinson <markw at illuminae.com>
Illuminae