[Gdpdm-devel] Ideas for future releases of GPDMD schema

[Ken Youens-C. <kc...@cs...>]

Terry,

As we discussed on this morning's Gramene conference call, here are a  
couple of ideas for possible inclusion in future GDPDM schema releases:

- Allow a many-to-many relationship between div_passport and  
cdv_source:  According to Isaak, germplasms may be available from  
more than one source.  Also, I would prefer to deprecate the  
"div_passport.source" field and move this info to the normalized  
"cdv_source" table.

- Allow for annotations/xrefs on germplasm names/synonyms:  This is a  
combination of ideas.  First off, it would be nice to attribute a  
germplasm's synonyms, e.g., to say that "IRGC 3575" is an IRRI  
synonym.  This could be accomplished by adding a "synonym_type" table  
with a FK to "div_synonym":

   CREATE TABLE div_synonym_type (
     div_synonym_type_id int PK,
     synonym_type varchar(50), -- e.g., "IRRI"
     xref_url text             -- e.g., sprintf-type URL template,
                               -- "http://www.iris.irri.org/action/ 
gms?method=getGmsList&search=%s"
   );

This would be even more useful if ALL germplasm names could be  
attributed and xref'd, so you'd remove the "div_passport.accename"  
and add a "div_display_synonym_id" field instead to indicate which  
synonym is used for the display.  It creates a somewhat bizarre  
circular reference (no worse than others already in GDPDM, like those  
that reference other records in the same table [div_parent_id]), but  
it makes for a very flexible way to both search just one table for  
germplasm names and to put their sources into the db.  It's possible  
some sort of "annotation" table could also be used, but this way just  
makes everything explicit.

ky