[Bioperl-l] Bio::DB:GFF

[Oliver Burren]

Hi Lincoln,

On Mon, 9 Sep 2002, Lincoln Stein wrote:

> This is just a stupid omission from the API (I come from a background where
> data is never deleted from a database, but merely marked with an "obsolete"
> flag).

This is how we usually work in our lab by pensioning off records rather
than deleting them. Is there any scope for adding 'date_from' and a
'date_to' field, so you can define a period that your data was 'live'
(present records have an explicit null in the date_to field)?


> I'm happy to fix, since you're not alone in requesting this feature.
> Can you suggest the form of the calls that you'd like to use to do deletes
> and updates?

Some ideas:

$gff_db->remove_seq($seqname) 	# Pension off all data contained within
				# this sequence


$gff_db->remove_features($seqname,
		[$method $source])	# Pension off all data which has a
					# particular method or source

$gff_db->update_all_seqFeatures($seq) 	#Sequence object including
					#features to dump
					#to the database

$gff_db->update_feature($fid,$feature)  #update a particular feature

I understand that others may want things done differently but, in our lab
we have our regions and we analyse them fairly atomically hence the lack
of coordinate arguments.

Let me know how/if I can be of any assistance, and thanks for ggb !

Olly

>
> Lincoln
>
> On Thursday 05 September 2002 09:01 am, Oliver Burren wrote:
> > Hi,
> >
> > I'm currently looking at the excellent Generic Genome Browser to visualise
> > some annotation we have in our lab. This has got me using Bio::DB::GFF
> > which works well for the addition of GFF files into a MySQL db. However
> > there seems to be no reciprocal provision for removing data from the
> > database (i.e. if it is wrong or need updating). Currently I just go in
> > and remove the data by hand using SQL. I was wondering if anyone has had
> > any thoughts about methods that allow you to remove data from database
> > using a Bio::DB:GFF object? For example 'remove all features contained in
> > segment IDDM1 with a source of CDS'? Do these methods exist already and I
> > just haven't found them ?
> >
> > Thankyou,
> >
> > Olly Burren
> > ---------------------------------------------------------------------------
> >---- JDRF/WT Diabetes and Inflammation Laboratory
> > Cambridge Institute for Medical Research
> > Addenbrooke's Hospital Site
> > Hills Road,
> > Cambridge
> > CB2 2XY                         Tel. 01223 762598 Fax. 01223 762102
> > ---------------------------------------------------------------------------
> >----
> >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l@bioperl.org
> > http://bioperl.org/mailman/listinfo/bioperl-l
>
> --
> Lincoln Stein
> lstein@cshl.org
>

-------------------------------------------------------------------------------
Computer Associate
JDRF/WT Diabetes and Inflammation Laboratory
Cambridge Institute for Medical Research
Addenbrooke's Hospital Site
Hills Road,
Cambridge
CB2 2XY                         Tel. 01223 762598 Fax. 01223 762102
-------------------------------------------------------------------------------