[Genex-users] Re: genex2 status?

[<ja...@op...>]

Michael Heuer <he...@ac...> writes:

> Could you comment a bit on the motivation or design rationale behind
> the genex2 schema/datamodel changes?  From what I've seen it looks
> like the team is going for MAGE compliance where it makes sense and
> cleaning things up where MAGE gets too verbose; as opposed to the
> ArrayExpress database which uses the entire model, warts and all.
> But that said, I could be way off...

That's pretty much it. ArrayExpress a complete one-to-one mapping of
MAGE whereas Genex was one of the DB's that MAGE was modelled
after. We've been adopting bits of MAGE as we've decided they make
sense.

Things we've adopted are:
* The BioAssay hierarchy: Physical -> Measured -> Derived
* The DesignElement Model: Feature, Reporter, CompositeSequence
* The Contact Model: Organization and Person
* The Experiment Model: including ExperimentFactor and FactorValue

Things we've modified:
* the Protocol model: we needed to support the ability for users to
  define computational protocols and run them against their data.
* Security model: MAGE doesn't really deal with this, Genex1 didn't
  either. Genex2 has full row-level security - so that if desired
  users can be given direct SQL access to the DB without comprimising
  security. 
* Biomaterial: we're currently using a simplified version with only
  limited support for OntologyEntry's.
* ArrayDesign: we're only supporting the minimal subset needed.

Hope that helps.

One thing to point out is that the Genex2 schema is very flexible. It
can be reconfigured at install time to add new pieces and much of the
system can be rebuilt automatically to work with the new schema
changes. 

Cheers,
jas.