On 8 Dec 2006, at 16:07, Brian Osborne wrote:
Arek,
At this point do you know what schemas are in your list of schemas to
try
with MartBuilder? You mentioned Chado...
The idea of MBuilder is that people can try _any_ schema and it should
be able to convert it to a BioMart compatible one without any
additional help from our side.
We have tested a few 3NF schemas so far and MBuilder certainly can do
the basic
transformation there and algorithm as it stands seems to be doing ok so
it should
already be useful for people.
What we are planning to do next is to 'train' it to do a _full_ Ensembl
transformation
(as this also requires some extra slightly weird stuff ..) and move it
to Ensembl mart production
We almost there and this should happen in jan sometime. This will be
followed by
adding a 'Runner' component so MBuilder can also run and monitor those
transforms.
Then we would like to move on and try more tricky schemas like
Reactome, ArrayExpress
and chado in order to extend the current algorithm to cope with those
as well and finally
do similar things with creating BioMart schema from XML (dtd, xml
schema)
At which point we would hope that this would cover pretty most all of
it's use cases.
a.
Thanks again,
Brian O.
On 12/8/06 5:38 AM, "Arek Kasprzyk" <[EMAIL PROTECTED]> wrote:
However, as we are now going to work intensively on
MartBuilder to train it with variety of schemas, it is likely that
chado will be tried again
------------------------------------------------------------------------
-------
Arek Kasprzyk
EMBL-European Bioinformatics Institute.
Wellcome Trust Genome Campus, Hinxton,
Cambridge CB10 1SD, UK.
Tel: +44-(0)1223-494606
Fax: +44-(0)1223-494468
------------------------------------------------------------------------
-------
