[Open-bio-l] Schema for genes & features & mappings to assemblies

[Keith James]

>>>>> "Matthew" == Matthew Pocock <matthew_pocock@yahoo.co.uk> writes:

[...]

    Matthew> If Oracle is your target database then you may well be
    Matthew> able to use the oracle tree-fetching queries to pull out
    Matthew> entire assembly trees in one hit rather than performing
    Matthew> multiple requests. Failing this, a little bit of
    Matthew> server-side script can be convinced to walk the tree for
    Matthew> you (if your db admin allows this kind of thing).

Or you can represent the trees as nested sets. You can get a subtree
in one hit then. The cost with this is that inserts/deletes of nodes
becomes quite complex as you need to tweak lots of surrounding nodes
(depending on how you store the nesting information).

There's people in our group (Ed Zuiderwijk and David Harper) working
on storage of assemblies in MySQL to circumvent the need to split GAP4
databases (as I understand it). Perhaps it's worth dropping them a
line?

Keith

-- 

-= Keith James - kdj@sanger.ac.uk - http://www.sanger.ac.uk/Users/kdj =-
Pathogen Sequencing Unit, Wellcome Trust Sanger Institute, Cambridge, UK