[BioSQL-l] Help with load_seqdatabase.pl

[Hilmar Lapp]

On Tuesday, January 28, 2003, at 01:55  PM, Jansen E Lim wrote:

>
> In addition to the VERSION attribute in bioentry, I would probably add 
> a STATUS attribute.  This
> attribute could have either of the following values 
> CURRENT|DELETED|HISTORY.  So, a bioentry
> record will have the bioentry_id (OID) as the primary key generated by 
> the system.  For a unique
> key, why not concatenate the ACCESSION and VERSION attributes? The 
> STATUS attribute will
> clearly tell us which is current.  A status of 'DELETED' just means 
> that in a "normal" query, that record
> won't come up.  So, we could have the following tuples in the db:
>
> OID    ACCESSION    VERSION    STATUS  ...     Comment
> 1        NM_001808    1        HISTORY            Original record 
> 2        NM_001808    2        HISTORY            spelling changed in 
> the description
> 3        NM_001808    3        CURRENT            sequence text is 5 
> bases longer
>

Right. As I mentioned you can bolt that on using the qualifier/value 
system. I'm not particularly in favor of incorporating it into the 
bioentry table, because 1) it pertains to LIMS and Biosql isn't 
targeted at LIMS anyway, and 2) it kind of defies the design of lean 
core tables with infinite extensibility through ontologies (which does 
come with a benefit, but also at a price).

The triple of (Accession,Version,Namespace) is indeed the compound UK 
(another one is Identifier).

	-hilmar
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------