[Bioperl-l] Bio::Ontology & Bio::OntologyIO

Hilmar Lapp hlapp at gnf.org
Thu Mar 6 17:32:54 EST 2003


I just committed an extensive test of the SO parser, added to 
simpleGOparser.t (as I said, GO and SO parsers use the same codebase, 
namely the dagflat parser). It did reveal a few problems where 
identifiers where checked to strictly for being beginning with 'GO:'. I 
corrected that.

All tests pass, indicating that the OntologyEngine that works behind it 
(Graph.pm-based SimpleGOEngine) is relatively (in fact surprisingly) 
flawless. Note that the test runs directly off of the latest SOFA from 
song.sf.net, which hereby becomes part of bioperl 
(t/data/sofa.ontology).

So, bioperl is ready for GO and SO from the framework perspective. 
Again, if you're considering using ontologies in bioperl, check out in 
particular the following interfaces: Bio::Ontology::OntologyI and 
Bio::Ontology::TermI. If you think there's an API function missing, 
speak up *now*.

	-hilmar

On Thursday, March 6, 2003, at 01:14  AM, Hilmar Lapp wrote:

> I've just committed another series of fixes and changes. This should 
> complete the Ontology module reorganization and clean-up.
>
> 	- moved more generally useful methods into an expanded TermI/Term; 
> GOterm is now a very thin layer on top of Term (basically, it provided 
> an alias GO_id to identifier() which format-checks)
>
> 	- the GO parser doesn't contain anything restricted to GO anymore; I 
> renamed it to a base parser dagflat, from which the goflat and soflat 
> parsers inherit. All they add is essentially a default for the term 
> factory and for the ontology name.
>
> 	- as just mentioned, there is now a SO parser, which still needs a 
> test written for it though (provided SO flat format is the same as GO 
> flat format, there shouldn't be any surprises):
>
> 		$ontio = Bio::OntologyIO->new(-format => 'so', # or 'soflat'
>                                       -file => 'sofa.ontology');
> 		$so = $ontio->next_ontology();
>
> I recommend that interested parties have a look at this and if 
> possible test-drive it, because it will go into the stable branch 
> within the next days probably.
>
> 	-hilmar
>
> On Tuesday, March 4, 2003, at 02:18  AM, Hilmar Lapp wrote:
>
>> I've committed a series of further fixes and standardization to the 
>> Bioperl ontology modules. My feeling is they are now ready for prime 
>> time use. Everyone who is interested in using ontologies in bioperl 
>> can start toying around with them - ideally within the next days to 
>> help smooth out the sharp edges before it gets back-ported over to 
>> the 1.2.x branch. Checkout Bio::Ontology::Ontology[I], 
>> Bio::Ontology::Term[I], Bio::Ontology::Relationship[I], and 
>> Bio::OntologyIO.
>>
>> What is still missing is
>>
>> 	- a SO parser (but now the GO parser should be SO-ready, so that a 
>> SO parser supposedly can be created with minimal effort from it)
>> 	- generalizing to the base Ontology::Term implementation a number of 
>> properties the Ontology::GOterm class adds, such that eventually 
>> Ontology::GOterm becomes a very thin derivation of Term (that either 
>> covers SOterm, too, or can be copied with minimal modifications to 
>> obtain the SOterm class)
>>
>> I'll address both within the next days.
>>
>> 	-hilmar
>>
>> -- 
>> -------------------------------------------------------------
>> Hilmar Lapp                            email: lapp at gnf.org
>> GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
>> -------------------------------------------------------------
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at bioperl.org
>> http://bioperl.org/mailman/listinfo/bioperl-l
>>
> -- 
> -------------------------------------------------------------
> Hilmar Lapp                            email: lapp at gnf.org
> GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
> -------------------------------------------------------------
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at bioperl.org
> http://bioperl.org/mailman/listinfo/bioperl-l
>
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------



More information about the Bioperl-l mailing list