On 14 Feb 2007, at 20:47, Vallejos, Andrew wrote:
Arek:
First I want to apologize for emailing you directly rather than using
the Biomart mailing lists. I have tried submitting email to the list
[email protected] , but it does not appear to be working.
Hi Andrew,
we had problems with spam on mart-dev in the past and that's why this
is a 'subscribers only' list now. I have subscribed you there so you
can post it now.
I also cc'ed your mail to mart-dev so people with more 'GO' expertise
can comment on this.
Second, we are currently still working with BioMart 0.4, and we have a
couple of issues. The first issue is that are getting duplicate
entries. Currently when I am exporting gene ids from our biomart I
can get anywhere from 1 to 5 (that’s the highest I’ve seen) ids for a
single entry in the main table (the gene ids are keys so there should
be no duplicates).
you need to tell me more about how you are exporting those genes. Do
you get duplicates if you just export from the main table ei you only
use attributes and/or fitlers from the main table?
or do you see duplicates only when you use dimension fitlers and/or
attributes?
The other issue is a bit more complicated. So we have gene ids in the
main table and go slim annotation in a dimension table. We want a
user to be able to filter using the slim annotation and then when the
user exports gene ids and go slim annotations they receive all slim
annotations relevant to the gene ids. Currently, if a user filters
for mitochondrion genes then when they export the go slim information
they only received mitochondrion (even though we know that the gene
ids returned have other slim annotation associated).
we tend to flatten go annotations so user receive all the go info
associated with a give gene. Could you try 'Ensembl gene' dataset
on our central server 'www.biomart.org/biomart/martview' with Go
annotation and see if this behaves correctly for you or do you see
there
the same problem?
a.
Again Thanks,
Andrew
--
Andrew Vallejos
Web Developer, Department of Physiology/BBC
Medical College of Wisconsin
8701 Watertown Plank Road
Milwaukee, WI, USA
Phone: (414) 456-4104
Email: [EMAIL PROTECTED]
------------------------------------------------------------------------
-------
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
------------------------------------------------------------------------
-------
