[XMLPipeDB-developer] yeast GO issues

[Kam D. <kda...@lm...>]

Hi,

You'll have seen in my previous message that MAPPFinder is stalling due to 
the hugeness of the GeneOntology, specifically, I think it is the 
GeneOntologyTree table.  After I sent the e-mail, I went and looked and it 
has 200,000 records instead of the ~38,000 that were in the 2006 database.

Obviously, this is not something we are going to be able to fix this 
semester, but I think that there will be a viable fix we can try to pursue 
with our research team next semester, which is:

GO has created something called "GO Slim" where they have removed a lot of 
the very specific child terms so that you only get the broad categories of 
GO.  We can try using the GO slim instead of the entire GO for our 
gdb.  The only hitch, of course, is that it is only provided in the OBO 
format, not the OBO-XML format.

There is a perl script called map2slim 
(http://search.cpan.org/~cmungall/go-perl/scripts/map2slim) that purports 
to take a gene associations file (like our GOA, but remains to be seen if 
it actually takes the GOA exact format) and re-map it to the GO Slim terms, 
so I think that is probably what we will need to do.  I'm guessing that it 
will take some rounds of testing before we are sure it is working properly 
for us.

So in the meantime, Bernie is going forward with analyzing the Arava data 
with the old gdb and hopefully Kenny and Don are making some headway on the 
gdb.

Cheers,
Kam