[Biojava-l] Implementation of Feature

David Huen smh1008@cus.cam.ac.uk
Tue, 4 Sep 2001 19:16:28 +0100 (BST)


On Tue, 4 Sep 2001 ialg@cin.ufpe.br wrote:

> 
>                 -> About the modeling of the class Feature, instead of use a 
> recursive structure where Feature implements a FeatureHolder, the FeatureHolder 
> could be like a Propreties that already exists in java.util? It's hold the tree 
> features but doesn't make necessary the recursive navigation, in the truth, all 
> information is there, full, so methods extending the Type Propreties could be 
> implemeted to get only specific information.
> 
I hope I've understood you correctly.  As I see it features are
necessarily hierarchical given that a feature can be an aggregate of other
(sub)features, e.g. exons being a subfeature of transcript.  The
FeatureHolder exists to make this possible.  And transcripts can be
(subfeatures) of a gene feature. There is no hardcoded model within
BioJava and the precise model you choose to implement may be quite
different from this but the hierarchy permits you to organise this as you
please.

Regards,
David Huen