[Bioperl-l] Bio::Ontology additions

[Allen Day]

Hilmar,

using this file:

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/obo/obo/ontology/anatomy/cell_type/cell.ontology

primary relationships are parsed correctly, but secondary relationships
are sometimes not.  for instance, for entry CL:0000441 follicle_stem_cell
"follicle stem cell @ISA@ animal_cell" is not correctly parsed, while
"follicle stem cell @ISA@ epithelial_fate_stem_cell" is.

Upon closer inspection i see that the problem might be that
"follicle_stem_cell" is in the incorrectly parsed relationship is referred
to as "follicle stem cell", a synonym of CL:0000441.

This is also a problem with "blastoderm_cell" sometimes being referred to
in relationships as its synonym "blastomere".

So... it really looks like this is a bug with creation relationships
between entities where one or more of the entities is referred to by its
synonym rather than primary name.

-Allen




On Wed, 2 Jul 2003, Hilmar Lapp wrote:

> Allen could you add a test case (possibly with sample data?) that 
> reveals the bug so that we can start fixing it.
> 
> 	-hilmar
> 
> On Sunday, June 22, 2003, at 03:28  PM, Allen Day wrote:
> 
> > done.  btw, i've introduced a bug in the parsing of secondary term
> > relationships when they use custom predicates.  from what i've seen 
> > this
> > doesn't affect the graph structure as the secondary relationships are
> > redundant, but it still needs fixin'...
> >
> > -allen
> >
> > On Sat, 21 Jun 2003, Hilmar Lapp wrote:
> >
> >>
> >> On Friday, June 20, 2003, at 11:37  AM, Allen Day wrote:
> >>
> >>> So... I have added two new methods to SimpleGOEngine to accomodate
> >>> this:
> >>>
> >>> add_relationship_type( $type_name, $ontology)
> >>> 	- adds new relationship types that don't already exist in the
> >>> 	ontology.  creates Bio::Ontology::RelationshipTypeI objects.
> >>> 	returns 1 on successful creation, undef on failure.
> >>>
> >>> get_relationship_type( $type_name)
> >>> 	- returns the previously created Bio::Ontology::RelationshipTypeI
> >>> 	object, or undef if there isn't one.
> >>>
> >>>
> >>
> >> It sounds like these could be generally useful (or even required). If
> >> so you may want to consider adding them to OntologyEngineI; generally
> >> speaking the parser may be supplied with an engine implementation
> >> different from SimpleGOEngine (although I believe right now the parser
> >> will throw up for other reasons if you do that ...).
> >>
> >> 	-hilmar
> >>
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l at portal.open-bio.org
> > http://portal.open-bio.org/mailman/listinfo/bioperl-l
> >
> >
>