[Bioperl-l] GFF and LocatableSeq refactoring

Chris Fields cjfields at illinois.edu
Sun Aug 16 02:49:25 UTC 2009


On Aug 15, 2009, at 4:07 PM, Mark A. Jensen wrote:

> I'm all for an attempt to split out phylogenetic stuff, it
> seems natural, and think in terms of a phylo package
> dependent upon a sequence package, and if necessary
> vice versa -- although if the Bio::Species - Bio::Tree::Node
> connection is relatively loose, perhaps we can refactor to
> make some attributes/methods optional features that carp
> when the phylo package is not installed. (Roles, anyone?)

I'm pretty sure they're linked very tightly (Species is-a Bio::Taxon  
is-a Bio::Tree::Node). This may be something Sendu needs to chime in  
on; he refactored much of that code prior to 1.5.2.

As a suggestion, maybe we can use a combined strategy: fall back to a  
very simple Bio::Species container class if a bioperl-phylo isn't  
installed, but utilize Bio::Taxon when it is.

> However, probably 1.6.x doesn't sound like the place to
> do that! I myself wouldn't have any problem waiting till
> 1.7 for 'official' Nexml support--but I hope Chase will chime
> in on that. What does Chris think?
> MAJ

Robert's suggestion of a separate distribution makes sense; it may be  
one avenue of slowly migrating out phylo-specific code into it's own  
distribution.  Not sure about calling it bioperl-phylo (which might be  
confused with Rutger's Bio::Phylo).

chris



More information about the Bioperl-l mailing list