[Biocorba-l] SeqFeature and Id

[Juha Muilu]

Hi Brad

Brad Chapman wrote:
> 
> Hi Juha;
> Nice to hear from you!

Nice to hear from you too! And thanks for the feedback!

> 
> > One question which has been bothering us for some time is that should
> > the SeqFeature has id? There are databases which have just features
> > without sequences..
> >
> > So, shall we make the SeqFeature (or even Annotation) inherit from the
> > Identifiable?
> 
> I'm indifferent about the change -- for me it doesn't make much
> sense to have a sequence feature without a sequence (hey, the
> sequence is the important part, right? :-).

That is right. In many cases the id is obsolete.

However there can be cases where you like to find/identify some certain
features  E.g. features/genes in the ensembl database. Also the id can
be a key to other database which provides more information about the
SeqFeature (or Annotation as well).... How do people feel about this?

> 
> What I'd like to see is a concert IDL we can implement -- see how it
> works and then adjust from there. This worked well for BioCorba: we
> had an initial spec that worked good for us, then Alan came along
> with the EMBL stuff and beefed it up for what he needed.

That is the way to go. 

> 
> > BTW I have made (eh already sometime ago) IDLs and clickable UML
> > documents out of the version we agreed during the BOSC. Those are on
> > page: http://industry.ebi.ac.uk/BSANE
> 
> Cool. Thanks. I'm almost finished with a Biopython implementation of
> the spec (hopefully after this weekend). I've got the Sequence and
> Annotation stuff done, almost finished with SeqFeatures, and then
> just collections to go. Yippee!

Really nice! I guess we need to get that EMBL wrapper up and running
asap :-).

> 
> Since we're talking about the spec, I might as well bring up a few
> snags I've got so far so people can comment on them:

Thanks a lot!!

> 
> --> The current idl won't compile with orbit-idl (so orbit-python
> and orbit-perl won't work with it). There is a bug in the IDL
> compiler (at least up to 0.5.8) and it can't handle recursive
> structs. So the SeqFeatureLocation struct, which has a reference to
> a list of SeqFeatureLocation sub_regions, will cause the compiler to
> segfault. I'm not sure what the best thing to do about this is.
> Suggestions from anyone?

I have reported the bug some time ago. I will check what is the status.

> 
> --> The inclusion of CORBA.idl causes lots of problems for me.
> First, CORBA.idl isn't included with any python ORB implementations
> as far as I can tell -- I found some in a Java ORB, but it didn't
> contain the definitions we are using (ObjectSeq, OctetSeq,
> StringSeq, ULongSeq). My preference would be to just add these
> definitions to the bsane.idl file and not worry about it -- what do
> people think about this?

The typedefs should be part of ORB, but it seems that there are no
standard way to include those. I think we should use existing
definitions and make the CORBA.idl interchangeable, or something. Any
other comments?

> 
> Other than that it's going okay. I'll post again here once I've
> finished the python implementation so anyone interesting can play
> with it.
> 
> Sorry for the long mail. Hope ya'll are having a good Friday.
> Brad

Very nice job Brad! These kind of mails keep us running (on right
direction).

Regards, Juha


> _______________________________________________
> Biocorba-l mailing list
> Biocorba-l@biocorba.org
> http://www.biocorba.org/mailman/listinfo/biocorba-l

-- 
 +--------------------------------------------------------------------+
 |Juha Muilu, Ph.D., EMBL Outstation| Email:  muilu@ebi.ac.uk         |
 |European Bioinformatics Institute | Phone:  +44 (0)1223 494 624     |
 |Wellcome Trust Genome Campus      | Fax:    +44 (0)1223 494 468     |
 |Hinxton, Cambridge CB10 1SD, UK   | http://industry.ebi.ac.uk/~muilu|
 +--------------------------------------------------------------------+