[Bioperl-l] Obsolete method call in Bio/DB/SQL/SeqAdaptor.pm?

Marc Logghe Marc.Logghe at devgen.com
Thu Feb 3 11:45:51 EST 2005

> The bioentry table has a key biodatabase_id that includes 
> biodatabase_id,
> accession, and entry version. The fasta files have only a 
> number on the
> header line to identify them and apparently either SeqIO or 
> Seq loads this
> into the display id and leaves the accession undefined so 
> that all of the
> records have the same key (e.g., 4-unknown-UNK) which is a 
> Bozo no-no (for
> you old enough to remember who Bozo is). I'm using bioperl 
> 1.4. Shouldn't
> the accession get set to display_id if accession doesn't 
> exist? Failing
> that, would it make sense to modify the store method in 
> SeqAdaptor to use
> the display_id in the key if the accession is undefined?

What I do in that case is use the -pipeline option of load_seqdatabase.pl and pass it a seqprocessor that sets the accession number to the display_id.
load_seqdatabase.pl <my_other_options> --format fasta --pipeline "SeqProcessor::Accession" <fastafile>

You create first the use-able SeqProcessor::Accession processor that basically should look like this:

package SeqProcessor::Accession;
use strict;
use vars qw(@ISA);
use Bio::Seq::BaseSeqProcessor;
use Bio::SeqFeature::Generic;

@ISA = qw(Bio::Seq::BaseSeqProcessor);

sub process_seq
  my ($self, $seq) = @_;
  return ($seq);


I am a big fan of the seqprocessor facility of Bioperl. It is just magic.
Everybody should make use of it ;-)

More information about the Bioperl-l mailing list