[Bioperl-l] 'virtual' seqs

Lincoln Stein lstein@cshl.org
Mon, 1 Jul 2002 13:14:53 -0400


Ah, left out the important part...

	I am happy with an EmptySeq that returns undef because I already am
	paranoid and check for seq() returning undef (since Bio::DB::GFF::Segment
	can do this).

	ADD: ... but I'm sure that not everybody shares my paranoia and returning
	undef will break some number of applications.

Lincoln

>
> Lincoln
>
> On Friday 28 June 2002 11:00 pm, Elia Stupka wrote:
> > > IMHO the connotation of UnknownSequence is overloaded too (refering to
> > > a sequence for which there is no functional annotation, or which
> > > hasn't been characterized experimentally yet). It appears that
> > > VirtualSeq has some consensus ... even though I recall someone (Elia?)
> > > saying that's too overloaded either.
> >
> > I do think that VirtualSeq (not only in Ensembl) is often used to
> > indicate a virtual *longer* sequence made by pulling together an assembly
> > of sequences.
> >
> > The term Lazy really has to do with lazy fetching and nothing to do with
> > an empty sequence.
> >
> > I still think EmptySeq is the simplest way to call something that we are
> > all referring to as an empty sequence.... ;) but I don't really see why
> > it should return NNNs... it is not meant to be storable with a sequence.
> > In other words there is a good reason why EmptySeq->seq should return
> > undef, and that is that it is not meant to be filled with a sequence, and
> > thus no sequence is meant to be stored...
> >
> > I totally agree with Hilmar, we have tons of fully masked sequences,
> > which are all made of NNNs... as well as gaps,etc. all made of NNNs, it
> > really is not the same thing as an EmptySeq...
> >
> > Elia

-- 
========================================================================
Lincoln D. Stein                           Cold Spring Harbor Laboratory
lstein@cshl.org			                  Cold Spring Harbor, NY
========================================================================