[Bioperl-l] removing Seq features

Mark Wilkinson mwilkinson@gene.pbi.nrc.ca
Wed, 18 Oct 2000 13:28:42 -0600

> has something to do with the supposed immutability of Seq objects, right
> Ewan? Otherwise it's a bug.

given that the science of studying these sequences is itself mutable...

but seriously - what are the arguments *against* this?  You are right in
saying that we *really* want to be able to add/drop features, at will, during
our analysis of any given sequence.  Lets hear it for on-the-fly manual review
and hand annotation!  Speaking as a molecular biologist, computer predictions
alone just don't cut the mustard!!..... >ahem<... oh, hello Ewan... um...
apologies to ENSEMBL  ;-)

At present, our WorkbenchII/SeqCanvas project doesn't add/remove features
to/from the Seq object it is displaying; since these features are, for us,
"live" database objects, we strip them out of the Seq object as it is loaded
up, then clone them and modify them as we wish without ever talking to "Seq"
again... but it would be more 'intuitive' if the graphical display was a live
display of the Seq object it is supposed to be representing... not only for
common sense reasons, but also because the user will almost certainly not want
to lose the changes to his/her precious Seq object, and would thus want to be
able to pull the entire thing out intact and up-to-date on his/her whim...

So, again, why should it *not* be this way... and would it be painful to make
it so?

Cheers all!


Dr. Mark Wilkinson
Bioinformatics Group
National Research Council of Canada
Plant Biotechnology Institute
110 Gymnasium Place
Saskatoon, SK