[Bioperl-l] bioperl-db Species question

Chris Fields cjfields at uiuc.edu
Wed Jan 31 11:23:52 EST 2007


Hilmar, Sendu,

I have a couple of quick bioperl-db questions; no hurry if you have  
your hands full.

I was working on fixing a bioperl-db bug for Roy, which is also  
related to Brian's bug:

http://bugzilla.open-bio.org/show_bug.cgi?id=2197
http://bugzilla.open-bio.org/show_bug.cgi?id=2092

Two things popped up:

1)  As Sendu alluded to in Brian's bug report (bug 2092),  
Bio::DB::BioSQL::SpeciesAdaptor has the current behavior of storing  
Bio::Species data by genus/species names (single name for each) via  
resetting classification(), which is inconsistent with Sendu's  
Bio::Species/Bio::Taxon changes.  Currently, Bio::Species stores  
names strictly by genus/species node names, where the species node  
name normally also contains the genus name.

I know where to make initial changes if we want to use node names in  
bioperl-db (SpeciesAdaptor::populate_from_row), but this currently  
break some bioperl-db species tests (02species.t).  I can also see  
this being a big headache, though I couldn't find anything that  
indicates doing so isn't BioSQL-compliant.  What would be the best  
way to go here?

2)  Somewhat related: should we be running strict node name checks  
using throw() in Bio::Species at all times?  There is no current  
workaround using FORCE, as is suggested in Bio::Species POD.  This  
currently seems to cause more problems than it's worth and sometimes  
throws even when running tax lookups with bioperl-db (Roy's bug).  I  
have changed the throw() to a warn() for the time being.

In the meantime, I am also converting all tests in bioperl-db to  
Test::More (I'm about 3/4 through now); I'll add a copy of  
Test::Simple to bioperl-db t/lib per Michael Schwern's instructions  
JIC it isn't installed.  Let me know if this is a problem.

cheers!

chris

Christopher Fields
Postdoctoral Researcher
Lab of Dr. Robert Switzer
Dept of Biochemistry
University of Illinois Urbana-Champaign





More information about the Bioperl-l mailing list