[BioSQL-l] Gene Cluster

[Hilmar Lapp]

Hi Raoul,

On Jan 21, 2008, at 7:23 AM, Raoul Jean Pierre Bonnal wrote:

> [...] Using bioentries is right if you have an entry which  
> represents a single
> cluser. Unfortunately I have to represent clusters inside "complete"
> genome, so a bioentry can have multiple clusters.

Note that relationships between bioentries are no different from  
those between seqfeatures. A bioentry, or seqfeature, can have  
relationships with any number of other bioentries and seqfeature,  
respectively. The uniqueness constraint only says that between any  
two bioentries, or seqfeatures, there can only be one relationship of  
one type (but multiple ones if the types are different).

>
>> Obviously, you can do the same for seqfeatures using
>> feature_relationship.
> Watching at BioSQL' schema, this seems to be te only solution for
> representing clusters in a general way.

See above as to whether this is the only way or not. The choice  
depends on what kind of objects you are trying to relate to each other.

> Now which is for you the most correct Feature Key misc_feature or
> source ?

Do you mean for type of the feature or type of the relationship? I'm  
a bit confused. For relating two features to each other it doesn't  
matter what type they are of; however, if they are sequence  
annotation features, you may want to make sure that the types and the  
hierarchy are compliant with the Sequence Ontology.

>
> I'm using BioRuby and working on it for interfacing it with BioSQL

Cool!

	-hilmar
-- 
===========================================================
: Hilmar Lapp  -:-  Durham, NC  -:-  hlapp at gmx dot net :
===========================================================