[Bioperl-l] suggestions for suitable 'taxon' object

Chris Fields cjfields at uiuc.edu
Wed Dec 20 01:16:47 EST 2006


On Dec 19, 2006, at 10:23 PM, Rutger Vos wrote:

> Hi all,
>
> I am looking for a bioperl object that can be abused to function as a
> suitable 'taxon' object, where I mean 'taxon' as understood by the  
> NEXUS
> file format (i.e. not strictly an entity from a taxonomy, but more  
> loosely
> an OTU).
>
> The object would primarily function as a way to relate nodes in  
> trees to
> sequences in an alignment (a foreign key that both nodes and  
> sequences refer
> to), and secondarily as the keeper of the canonical name of the  
> OTU, such
> that a sequence named 'Homo_sapiens|EF177447.1/12-56' and a node  
> named 'Homo
> sapiens (constrained monophyly)' can still be understood to refer  
> to the
> same thing - the OTU 'Homo sapiens sapiens' (for example).

Alignment (SimpleAlign) objects contain Bio::LocatableSeq sequence  
objects; at the moment LocatableSeqs don't store their own annotation  
but they could easily be made or subclassed to be AnnotatableI (i.e.  
they can store annotation collections).  I recently made SimpleAlign  
Annotatable; Jason has also made SimpleAlign implement  
FeatureHolderI, so alignments can store SeqFeatures as well; he may  
have his own designs here.

There may be a wide variety of ways to go about this.  I would  
probably do the following (bear in mind I'm a microbiologist, not a  
computer scientist).  If one could add trees as annotation to the  
alignment (i.e. if trees could be Annotation objects and kept in the  
SimpleAlign's annotation collection), and each sequence in the  
alignment contained reference to a node object of the tree (i.e. if  
Bio::Taxon/Bio::Species objects could also be Annotation objects, but  
kept in a LocatableSeq annotation collection), both could refer to  
the same node object.  This may not be exactly what you want, but  
maybe it's close?

SimpleAlign->AnnoColln->Tree->OTU(Nodes)
    \----->LocSeqs-->AnnoColln-----/

I suppose this could also be done with Seqfeatures...

> I was thinking that a (possibly expanded) Bio::Species might work  
> if there
> was some sensible way of appending references to node and sequence  
> objects
> to it (or otherwise associate them with each other), but I am  
> writing *to
> solicit any and all suggestions*. I am looking for something  
> similar to
> Bio::Phylo::Taxa::Taxon.
>
> Any and all comments and suggestions greatly appreciated!
>
> Best wishes,
>
> Rutger Vos

Sendu would be the best one to speak about Bio::Taxon and  
Bio::Species and may have some ideas on the above.  The current plan  
was to deprecate Bio::Species, but who knows?

chris


More information about the Bioperl-l mailing list