[Biocorba-l] bioCORBA, BSANE SeqFeatures/Locations

[Juha Muilu]

Hi Jason,

Jason Stajich wrote:
> 
> Not sure where to start responding, so much traffic!

Welcome to the discussion !

> 
> I assume we are to be looking at
> http://corba.ebi.ac.uk/~muilu/uml/bsane_v3/idl/seqcore.idl to be the
> merger between BSANE and biocorba core objects and when agreed upon, the
> biocorba 0.3 spec?

Yes seqcore is the merger. The bsane.idl contains some general stuff
needed by the seqcore.

> 
> Very happy to hear that Vectors may die!  I found I had to do very
> stupid things in bioperl/ensembl corba bindings to comply with this
> interface and then Alan told me he would just not implement methods he
> thought were uncessary...  Gosh, I should have learned.
> 
> SeqFeatures:
> 
> I agree with Ewan's comments on what is missing for GFF.  

Sorry. I missed those. The bioCORBA 0.3.1 (from Alan) has:

	// Return the source of the Annotation. This is included
	// mainly for GFF compatibility.
	string source();         

	// Return the primary ID of the Seq that this Annotation
	// belongs to.
	Identifier seq_primary_id(); 

Are these missing the bits? Any more? Methods are in the Annotation
interface in this BioCORBA version. Shall we keep those in that
interface or should we move them to SeqFeature. They may be for more
general use, so the Annotation may be OK place.

Should we replace the seq_primary_id by a method for getting reference
to owner sequence as proposed by Ewan
(http://biocorba.org/pipermail/biocorba-l/2001-June/000233.html)

> I notice a lot
> of "this is what biojava does", do we need to help explain what bioperl
> does better to make sure all those views are at least heard?

Sorry about the bias! I am just more familiar with the bioJava (not any
expert on that either). 
I would be happy to know where one solution works better or then fails.
bioCORBA/BSANE is used to glue the solutions together so it should
contains best parts of all, if possible

> 
> SeqFeatureLocations:
> 
> Glad you put region_operator in there, that will help handle a set of
> cases that were not possible with the existing biocorba spec.
> 
> I do think that 'sub_regions' or whatever we decide to name here will
> be used to have Locations which contain locations even though we did not
> really handle this in biocorba 0.2.  Implementations in bioperl and I know
> biojava do support hierarchies of locations, it just didn't bubble to the
> surface in time to be part of the 0.2 spec (Brad/Jeff I'm not sure if you
> guys got to this yet).
> 
> Additionally in bioperl we have a pluggable interface for how to
> interpret locations, because it is a little unclear how you plan to
> resolve the start() or end() calls on a SeqFeature with a set of
> FuzzyLocations.  We have a coordinate policy object which one can plug
> into the location object - Narrowest, Widest, and Average policies are
> implemented so that when one calls start or end() on a Location which is
> fuzzy or contains sub locations it can make a guess as to what you expect
> here.

That is nice solution. We need this as well. Already, to be able to
deliver results which uses the policies. Should the SeqLocation have
reference to the  LocationPolicy object?


> 
> This may be over the top for what is desired in the seqcore proposal, but
> at least explaining what we've done in bioperl.
> 
> -Jason
> 
> Jason Stajich
> jason@chg.mc.duke.edu
> Center for Human Genetics
> Duke University Medical Center
> http://www.chg.duke.edu/
> 
> _______________________________________________
> 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|
 +--------------------------------------------------------------------+