On 2 Oct 2007, at 20:56, Alessandra Bilardi wrote:
Hi all,
I test Biomart .5 over Chado.
I installed Chado following
http://gmod.cvs.sourceforge.net/*checkout*/gmod/schema/chado/
INSTALL.Chado without problems.
I loaded into Chado
ftp://ftp.yeastgenome.org/pub/yeast/data_download/chromosomal_feature/
saccharomyces_cerevisiae.gff sample. My database is sgdlite.
I created with MartBuilder a Schema about sgdlite and I created a
Dataset with only feature table of schema Chado of sgdlite database. I
created SQL script and I runned the SQL script. I had problem about
"No space left on device".
I checked free disk space for script running and I discovered that
when the script creates one relation, it uses over 20Gb so it keeps
busy all free disk space..
I don't think it is normal, so I suppose that I mistake when I create
the dataset..
I would to create one dataset that it allows me to ask biomart complex
query about sgdlite database. Complex query about genes, annotation,
ontologies, and other element that I will load in the feature table.
But I don't understand how to create dataset..
Can someone help me?
Thanks very much for the help in advance.
Hi Alessandra,
There is not yet an automated (and robust) process of a complete
conversion of chado schema into a biomart compatible schema.
However you may want to have a look at Don Gilbert's GMOD Tool
gff2biomart.pl to start with. You can also use MBuilder
but with limited success at the moment as MBuilder has been trained
against a 'standard' 3NF schema and does 'untangles'
feature->feature_relationship correctly as yet. We are currently
working on implementing this type of functionality but I must
admit that this has been lagging behind a bit due to us being rather
understaffed at the moment. Having said that
MBuilder should be fully capable of creating a simple dataset based on
the feature table without a problem and running
out of space. This sounds more like a bug to me. Would you be able to
give us your database dump so we could have
a look and recreate the problem here?
Finally there is always a possibility by manually achieving the
conversion chado->biomart by writing your own conversion
SQL (this how this is has been mostly done so far eg in ensembl) to
produce a correct transformation. I know that some
people on this list attempted to do it but I am not sure how successful
this was.
I would also advise you to move to release 0.6 as 0.5 tools are rather
out of date now
sorry could not be of more help,
a.
Alessandra.
------------------------------------------------------------------------
-------
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
------------------------------------------------------------------------
-------
