[Bioperl-l] Re: Seq or SeqIO to BSML

[Ewan Birney]

On Fri, 3 Aug 2001, Nathan O. Siemers wrote:

> 
> 
> So if we generate an ensembl embl format dump and then load it into
> bioperl via seqio, that will be equivalent for our purposes to using
> the original ensembl Seq(IO) object?  That's fine with me, but I
> thought you had voiced a reservation to doing it this way in
> Copenhagen. Perhaps I misunderstood.  The eventual point will be to
> create a piece of code that will work with ensembl's native tongue
> (SeqIO I assume) and send out a stream of xml/bsml from the ensembl
> site.

It will be a start, and you can get the basic structure of BSML right.
Once the structure is there, I'll take a look at it and see if I can add
in Ensembl specific parts to make it play well with Ensembl.


You could develop directly against Ensembl. The steps would be:

either (a) download database files or (b) make sure your Perl DBI can see
kaka.sanger.ac.uk

checkout from anonymous cvs from ensembl the 1.1 branch

use a script which builds a virtualcontig (there are examples in
doc/ensembl.tex) and use that to dump


but I suspect this will be too much work. I would play to our repsective
strengths - you develop the basic SeqIO dumper and I'll provide the
Ensembl specific elaborations.


ewan