[DAS2] URIs for sequence identifiers

[Andrew Dalke]

Ed:
>> Problem #2: feature searches return features from either namespace
>
> A different solution:
>
> Scaffold and Chromosome coordinate systems are served by separate 
> DAS/2 servers.  Each server returns data from one and only one 
> namespace.
>
> Those separate servers can, behind-the-scenes, use the same database.
>
> DAS/2 clients, like IGB, would choose to connect to either the 
> Scaffold-based server or the Chromosome-based server, but not usually 
> to both at once.
>
> Does this handle all the issues?

Here's the email I got from Andreas when I proposed this.


>>> There may be more than one COORDINATE element if ... (XXX why?)
>
> There are several servers that understand more than one coordinate 
> system and
> can return the same type of data in different coordinates. (depending 
> on which type of accession code/range was used for the request )
> E.g. there are a couple of zebrafish servers that speak both  in 
> Chromosome and Scaffold coordinates.
> (reason perhaps  being that zebrafish is an organism that seems to be 
> very difficult to assemble ?)


>> Will there be separate CAPABILITY items for each source?
>
> no. if there are then this should be registered as two independent 
> servers.

(but see clarification later)

> Allowing multiple coordinate systems per server is a way to slightly 
> reduce the already long list of known
> servers. Currently there are about 90 in the registry (+10 in the last 
> few weeks...) and there still are about 20 more
>  which have not been registered (and are provided by the BioSapiens 
> project).

>> Long for who?  It isn't that much data.
>
> It is long for somebody who browses manually through the ensembl DAS 
> configuration and searches for a DAS source to add to.
> It a "long" list for myself to read through the DAS server list at
> http://das.sanger.ac.uk/registry/listServices.jsp
> and although I know this list pretty well,  it seems to me a lot of 
> text/descriptions, etc.


>> There is only one reference server for an annotation server.
>
> I think it should be one reference server per coordinate system.


>> But if there are two COORDINATES elements, and you say that
>> each has its own reference server, then aren't you saying that
>> a single annotation server may have multiple reference servers?
>
> yes. i believe that this should be possible.

>>  What's the concern about having
>> no more than one coordinate per data source?
>
> Just last friday somebody asked me how to add a DAS server that has 
> two coordinate systems to different Ensembl views ( ContigView and 
> GeneView)
> Her initial solution was to provide multiple DAS sources
> http://das.sanger.ac.uk/registry/showdetails.jsp?auto_id=DS_211
> and
> http://das.sanger.ac.uk/registry/showdetails.jsp?auto_id=DS_219
>
> but I think this could be joint into a single server.


In any case, I think the proposal I outlined in the previous email
makes things cleaner even without support for multiple coordinate
systems on the same server.

					Andrew
					dalke at dalkescientific.com