[ensembl-dev] Projecting features

Tue Jan 15 14:43:20 GMT 2013

When I get a free year and loads of inspiration to reimplement the whole
API, these sort of things will be intuitive to do and beautiful to
behold.

I just need a bit of time... :-)

A

On Tue, Jan 15, 2013 at 02:37:21PM +0000, Michael Paulini wrote:
> Hi Dan,
> 
> have a look in ensembl-pipeline and and ensembl-analyis, there is some
> utility code to change some properties of Database objects there (like
> killing genes and having it cascade through associated things).
> 
> M
> 
> On 15/01/13 14:33, Dan Bolser wrote:
> > Many thanks Daniel,
> >
> > Following through this example, I thought I had found my error, namely
> > the update method needs to be called on the adaptor, not the feature
> > object... Unfortunately, I now get more bugs, but I'll try to resolve
> > them.
> >
> >
> > Thanks all,
> > Dan.
> >
> > On 15 January 2013 14:13, Daniel Barrell <db8 at sanger.ac.uk> wrote:
> >> Hi Dan,
> >>
> >> You could look at the transform example in:
> >>
> >> /ensembl-analysis/scripts/genebuild/copy_genes.pl
> >>
> >> I've used this script in the past to transfer genes between coordinate
> >> systems.
> >>
> >> Daniel
> >>
> >>
> >>
> >> On 15/01/13 13:54, Dan Bolser wrote:
> >>> Have you got code examples?
> >>>
> >>> Currently I may just dump TSV for SQL UPDATE...
> >>>
> >>>
> >>> On 15 January 2013 13:43, Andreas Kusalananda Kähäri<ak4 at sanger.ac.uk>
> >>> wrote:
> >>>> Hi,
> >>>>
> >>>> During the gene build process we regularly modify built models in bulk,
> >>>> but never using update() AFAIK.  Instead the results of a pipeline step
> >>>> is written to a new (empty) database.  The final database that we hand
> >>>> over for the release is created by copying data from a few of these
> >>>> result databases.  This way it's easy to go back to a previous step and
> >>>> rerun from there with modified parameters/code.
> >>>>
> >>>> Cheers,
> >>>> Andreas
> >>>>
> >>>>
> >>>> On Tue, Jan 15, 2013 at 01:03:12PM +0000, Kieron Taylor wrote:
> >>>>> Hi Dan, this is trickier than you might like.
> >>>>>
> >>>>> In general, database changes are made in response to ->store() or
> >>>>> ->update() methods on adaptors. update() calls may not change every aspect
> >>>>> of the object. I suspect update() will not have the desired effect on all
> >>>>> Features, but you should investigate this yourself.
> >>>>>
> >>>>> The problem with using store is the issue of storing the same Feature
> >>>>> again, but with new coordinates. This will probably trigger parts of the
> >>>>> code that prevent the storing of the same Feature twice, or else give you
> >>>>> two copies with slight differences. You may need to resort to ->remove(),
> >>>>> and the internal db identifiers will change...
> >>>>>
> >>>>> In summary, tread lightly and verify your results before you unleash the
> >>>>> script.
> >>>>>
> >>>>> Kieron
> >>>>>
> >>>>> On 15 Jan 2013, at 11:36, Dan Bolser wrote:
> >>>>>
> >>>>>> I wrote a script to project features from contigs to chromosomes using
> >>>>>> the transform method.
> >>>>>>
> >>>>>> https://gist.github.com/4537997
> >>>>>>
> >>>>>>
> >>>>>> Once I have the projected feature, I try to update the database by
> >>>>>> calling (line 73):
> >>>>>>
> >>>>>> $ff->slice( $ffp->slice );
> >>>>>>
> >>>>>>
> >>>>>> however, this doesn't update the database with the new feature
> >>>>>> coordinates.
> >>>>>>
> >>>>>> Is there a method I need to call to update the feature in the database?
> >>>>>>
> >>>>>>
> >>>>>> Cheers,
> >>>>>> Dan.
> >>>>>>
> >>>>>> _______________________________________________
> >>>>>> Dev mailing list    Dev at ensembl.org
> >>>>>> Posting guidelines and subscribe/unsubscribe info:
> >>>>>> http://lists.ensembl.org/mailman/listinfo/dev
> >>>>>> Ensembl Blog: http://www.ensembl.info/
> >>>>> Kieron Taylor PhD.
> >>>>> Ensembl Core software developer
> >>>>>
> >>>>> EMBL - European Bioinformatics Institute
> >>>>>
> >>>>>
> >>>>>
> >>>>>
> >>>>> _______________________________________________
> >>>>> Dev mailing list    Dev at ensembl.org
> >>>>> Posting guidelines and subscribe/unsubscribe info:
> >>>>> http://lists.ensembl.org/mailman/listinfo/dev
> >>>>> Ensembl Blog: http://www.ensembl.info/
> >>>>>
> >>>> --
> >>>> Andreas Kusalananda Kähäri
> >>>> Ensembl Gene Annotation Team
> >>>>
> >>>> Tap tap tap.
> >>>>
> >>>> _______________________________________________
> >>>> Dev mailing list    Dev at ensembl.org
> >>>> Posting guidelines and subscribe/unsubscribe info:
> >>>> http://lists.ensembl.org/mailman/listinfo/dev
> >>>> Ensembl Blog: http://www.ensembl.info/
> >>> _______________________________________________
> >>> Dev mailing list    Dev at ensembl.org
> >>> Posting guidelines and subscribe/unsubscribe info:
> >>> http://lists.ensembl.org/mailman/listinfo/dev
> >>> Ensembl Blog: http://www.ensembl.info/
> >>
> >>
> >> _______________________________________________
> >> Dev mailing list    Dev at ensembl.org
> >> Posting guidelines and subscribe/unsubscribe info:
> >> http://lists.ensembl.org/mailman/listinfo/dev
> >> Ensembl Blog: http://www.ensembl.info/
> > _______________________________________________
> > Dev mailing list    Dev at ensembl.org
> > Posting guidelines and subscribe/unsubscribe info: http://lists.ensembl.org/mailman/listinfo/dev
> > Ensembl Blog: http://www.ensembl.info/
> 
> 
> _______________________________________________
> Dev mailing list    Dev at ensembl.org
> Posting guidelines and subscribe/unsubscribe info: http://lists.ensembl.org/mailman/listinfo/dev
> Ensembl Blog: http://www.ensembl.info/
> 

-- 
Andreas Kusalananda Kähäri
Ensembl Gene Annotation Team

Tap tap tap.