[Bioperl-l] t/SimpleAlign: not ok 18

Jason Stajich jason@cgt.mc.duke.edu
Mon, 16 Sep 2002 20:05:06 -0400 (EDT)


Allen Smith wrote:
> Well, I've been looking at doing the above - using bioperl-1.1.0 from a
> mirror site, since CVS was inaccessible (I've now downloaded bioperl-live
> HEAD). I'm currently looking at consensus et al stuff in SimpleAlign with a
> view toward enabling a loose consensus mechanism, with that also being used
> for tiebreaking in ties for strict consensus situations. This would be via
> matrices, which is leading me into splitting out the matrix code in
> Bio::Variation::AAChange and SimpleAlign (CONSERVATION_GROUPS) into a
> seperate Bio::Matrix (and Bio::MatrixI) module. Thoughts?

We'd like to see this done - but we also have distance matricies that we'd
like to be able to represent so we have to be careful with the
Bio::MatrixI interface.  If it can handle both substitution and
distance than I am fine - otherwise we may want to be explicit in the
names, I'm thinking like Bio::Matix::SubstitutionI Bio::Matrix::DistanceI
or something more clever.

I would definitely like to see the CONSERVATION_GROUPS moved into
something more configurable than the basic groups that I threw in there.

Also definitely agree on what your observations for the proper handling of
consensus generation.  Appreciate your willingness to lend a hand
and jump in there to fix things.   Need all the hands we can get!



-- 
Jason Stajich
Duke University
jason at cgt.mc.duke.edu