[DAS2] on local and global ids

Andrew Dalke dalke at dalkescientific.com
Thu Mar 16 16:06:26 UTC 2006

> I think it will help considerably to have a document that lists the 
> valid
> sequence IDs for popular annotation targets. I've spoken with Ewan on 
> this,
> and Ensembl will generate a list of IDs for all vertebrate builds. 
> I'll take
> responsibility for creating IDs for budding yeast, two nematodes and 12
> flies.

What should people use if there aren't defined?  Like now?

If everyone must use the same well-defined global id for the features
response then doesn't that mean we can't have any DAS servers
until this document is made?

Is the general requirement that the first person to make a server for
a given build/genome/etc. is the one who gets to define the
global ids?  Or is it Andreas at Sanger who defines the names?

Suppose one group in California starts defining names for, say,
the barley genome.  Another group in say, Germany, is also working
on the barley genome.  They hate each others guts and don't work
together, so they make their own names.  The names refer to the
same thing because it was a group in Japan which produced the
genome.  Do we wait for an alignment service?  An identity service?
before people can merge data from these two groups?

Maybe we can solve all this by having an identity mapper format.
And defer defining that format until there is a problem.

There is no perfect solution.  This is a sociological problem.

Gregg's current client, I think, used hard-coded knowledge about the
mapping between the two current servers.  Then again, his code
already supports a synonym table.

					dalke at dalkescientific.com

More information about the DAS2 mailing list