[Bioperl-l] [Obo-discuss] software developer resources, OBO API?

Erick Antezana erant at psb.ugent.be
Fri Aug 15 12:25:59 UTC 2008


Hi Hilmar,

Hilmar Lapp wrote:
> Hi Erick,
>
> how did you determine that go-perl is specific to GO? I've found it to 
> work quite well for any kind of OBO-formatted ontology.
we have used go-perl during the gestation (2005) of the ontologies we 
built/handled/etc. In particular while dealing with GO (as it was 
originally conceived for that purpose). It is extremely useful indeed. 
However; as new extensions were needed by our team, new modules were 
developed on top of the existing ones and that code was actually 
inducing a new development branch which ended up in something called 
onto-perl. Onto-perl, as you might have realized, is strongly influenced 
by go-perl. On the other hand,  we had had some communications with 
Chris Mungall at that time, and he manifested he might drop further 
development on go-perl. Nevertheless, some time later he told me he will 
continue, which is good since many systems are based on it. But, by then 
we already had a sort of independent package which has shown to be useful.
>
> Also, you note that BioPerl doesn't have the ability to write in 
> certain formats, and to intersect and "unify" (would you mind 
> explaining what you mean by that?) ontologies.
While working with several OBO ontologies, we needed to have them (or 
part of them) merged, intersected, join (=get one ontology=unify). It 
can be of course a bit subjective..since you can "unify" ontologies 
based on different features/approaches/etc and while building 
application ontologies (such as CCO) you might be confronted to identify 
identical terms coming from different ontologies and get only one in 
your integrated resource....
> It seems that your implementation of RDF etc export isn't really 
> reusable or modular in any way, 
The exports (RDF, OWL, ...) are part of the Ontology module's 
functionality. have you had any particular problems while exporting an 
ontology? Please let us know so that it can be fixed or improved. On the 
other hand, that module is nowadays undergoing a deep improvement (not 
released yet) to accommodate a huge set of "ontologiz-ed" resources into 
an RDF repository. I would be also interested in discussing about it so 
that we could improve it.
> but I'd love to bring the intersection function over to BioPerl 
excellent! I think this is an appropriate time to make the diverse 
developments converge so that the users community could have a standard 
set of tools. We are  interested in following up these discussions.
> (BTW when you decided to roll your own ontology API, did you get the 
> impression that BioPerl isn't receptive to you adding to it?).
as I mentioned, we took originally go-perl since it offered much more 
functionalities that BioPerl::Ontology::*
> Would you mind pointing me to the place in the code where I would find 
> that, as I can't seem to find it.
I forwarded your request to the developer who might give you more 
details about it.
>
>     -hilmar
>


cheers,
Erick



More information about the Bioperl-l mailing list