[BioSQL-l] KW field from Swiss-Prot in Biosql schema missing

[Hilmar Lapp]

On Tuesday, March 9, 2004, at 01:07  AM, Raphael A. Bauer wrote:

> That's not the case in my biosql Schema. Only the GN (Gene Name) Lines 
> from Swiss-Prot are parsed into Bioentry_qualifier_value.
> The KW Field is still missing.
>

You will need at least the latest stable release of bioperl, which is 
1.4.x The 1.2 series stores the keyword in a way that will not get 
persisted by bioperl-db.

To verify, read in a swissprot-formatted sequence with keywords and 
then print the 'keyword'-tagged annotations:

	$seqio = Bio::SeqIO->new(-format => "swiss", -file => 
">myswissprot.dat");
	$seq = $seqio->next_seq();
	foreach my $kw ($seq->annotation->get_Annotations("keyword")) {
		print "KW: ",$kw->value,"\n";
	}

This should print you all keywords that the sequence had in the input 
file.

The trick about serializing keywords to biosql is that there is no 
trick. Bio::Seq::RichSeq (instances of which are returned by the 
swissprot parser) stores keywords as Bio::Annotation::SimpleValue 
objects behind the scenes (since 1.4.x), and bioperl-db just serializes 
all annotation objects.

With bioperl 1.2 this will not work because that one stores keywords in 
a separate property.

	-hilmar
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------