[DAS] Finding the ADAM2 Gene via Ensembl DAS

[Steve Chervitz]

Lincoln,

On Sunday, Jan 26, 2003, at 15:31 US/Pacific, Lincoln Stein wrote:

> To find all the ADAM genes in gbrowse, type ADAM* into the search box 
> and hit
> <enter>.
>
> BTW, I'm having a mini DAS-hackathon this week at CSHL to fix up the 
> protocol
> so that named features can be retrieved in a more sensible way.  We've 
> got
> gbrowse running on top of DAS, but would like the search interface to 
> work as
> well on top of DAS as it does on top of a GFF database.  It annoys me 
> no end
> that searching for ADAM2 returns all the exons as well as the genes.

Were there any conclusions or new specs resulting from your 
DAS-hackathon?

In addition to named features, it would also be useful to run feature 
queries without specifying start,end coordinates on the sequence. For 
example, "get all features of type 'gene' on the sequence named 
'contig123'." You could do this by allowing a segment specifier without 
start,end ('segment=REF'). As an optional extension, REF could contain 
wildcard characters (leading to the degenerate segment=*).

Another suggestion: Allow the returned features to contain arbitrary 
tag=values, a la the attributes in your GFF3 proposal. This would also 
necessitate having a way to specify in the request what attributes 
you're interested in getting back.

Regarding the GFF3 proposal 
(http://bioperl.org/pipermail/bioperl-l/2003-February/011191.html), 
have you considered writing a DTD for GFF3 and using that as a basis 
for DAS2? Just a thought.

Steve

PS. Gregg will be at CSHL for the upcoming Genome Informatics meeting 
and is eager to talk to you about DAS stuff. Keep me posted.