[DAS] pathways

[Chris Mungall]

On Wed, 17 Apr 2002, Kulp, David wrote:

> I haven't looked at AGAVE for a while, but my biggest concern is that it
> does not have a good implementation of alignments/assemblies.  Alignments,
> for example, were stored as BLAST snippets, if I remember correctly.

GAME XML handles this rather nicely.

...unfortunately there's no decent documentation on GAME at this time.

> My gut feeling is that ALL of this boils down to a clean representation of
> alignments.  This becomes a key issue when dealing with genome assemblies
> and versioning.  One can imagine, for example, an EST aligned to a BAC, the
> BAC assembled to a contig for version A, and that contig aligned to a new
> genomic assembly contig for version B.  If alignments are fundamental, then
> transfering annotations becomes much more tractable.

yup, all you need are feature and feature pair classes with a graph for
the compositional relationships.

what about tightly coupling the xml to the bioSQL relational schema?

> Will DoubleTwist's bankruptcy impact this decision?  The agave web site is
> down and DoubleTwist owns the DTD and has, presumably, the right to modify
> the licensing terms at any time.
>
> Besides these concerns, AGAVE was pretty nice.
>
> -d