[DAS2] Sanger/EBI trip report

[Helt,Gregg]

> -----Original Message-----
> From: das2-bounces at portal.open-bio.org
[mailto:das2-bounces at portal.open-
> bio.org] On Behalf Of Andreas Prlic
> Sent: Wednesday, October 19, 2005 1:56 AM
> To: Andrew Dalke
> Cc: DAS/2
> Subject: Re: [DAS2] Sanger/EBI trip report
> 
> Hi Andrew!
> 
> Thanks for the very good summary.
> 
> I wanted to comment on the DAS -registry we are having here at
> http://das.sanger.ac.uk/registry/
> 
> So far there has not been a mechanism how DAS clients can
> programmatically discover available DAS sources, that can be
> shared between different DAS clients. This registry addresses
> that issue (among others).

As part of the work under the DAS/2 grant, we've had several meetings
trying to sketch out a strategy for registry/discovery of DAS/2 servers.
Andreas presented his DAS registry at one of these back in September
2004.  Looks very nice.  There's still a lot of differing opinions about
the best way to do this though.  Maybe in November we can devote at
least one of the DAS/2 teleconference calls to revisiting
registry/discovery strategies, if Andreas could join in.
 
> One thing that was necessary to do for this, was to provide
> "coordinate systems", that describe the data that has been
> annotated. E.g. Ensembl can project data from different
> coordinate systems into one display so it is important to know
> what the data is being served in. - also some coordinate
> systems are not supported and therefore such DAS sources have
> to be ignored.
> 
> What I understood from our discussion there is no convention
> on globally unique coordinate systems in DAS2 so far, but I think
> that would be a nice feature to have.
> 
> There is some documentation on the registry available at
> http://das.sanger.ac.uk/registry/help_index.jsp
> 
> Greetings,
> Andreas
>

As far as globally unique coordinate systems, the current plan in DAS/2
is that this should be indicated as a URI id attribute in an <ASSEMBLY>
element for the versioned source.  For example, as part of the response
to http://server/das/genome:

<SOURCE id="human" ...>
	<VERSION id="Human_May_2004" ...>
		<ASSEMBLY
id=http://hgdownload.cse.ucsc.edu/goldenPath/hg17/" />
	</VERSION>
</SOURCE> 

and allowing multiple <ASSEMBLY> tags to indicate different URIs for the
same genome assembly.  

Preferably the assembly id would be a URI controlled by the institution
that performed the assembly, and be resolvable to some meaningful
information about that assembly.  In practice that might not be possible
(for example the above snippet refers to a download page at UCSC but the
assembly was done at NCBI).

	gregg