[Bioperl-l] Re: Bio::Tools::Phylo::PAML::Codeml

[Jason Stajich]

On Tue, 18 Jun 2002, Aaron J Mackey wrote:

>
> Jason, I've committed my first-pass attempt at an API by example in the
> SYNOPSIS section.  Let me know what everyone likes and dislikes.  I'm not
> (yet) making use of the Bio::AnalysisParserI or AnalysisResultI - are
> these stable interfaces to make use of (if so, then the


> $result->next_feature() doesn't make much sense for PAML) ?  Is there some
> other Bio::Factory::Bleah module that I need to be paying homage to?
>

I guess we could require the Factory::TreeFactoryI since we can produce
trees from the PAML objects - that would be my only suggestion there.  But
there aren't any other interfaces other than the AnalysisResultI which I
am not sure has shown much utillity at this point.  We can sort that out
later I think.

As for the examples in the SYNOPSIS - I think it is a great - I would
suggest a couple of things to smooth the transition to more parsers for
PAML output.

# parser will be a Bio::Tools::Phylo::PAML::IO::codeml object.
# ala Bio::SeqIO::genbank -- I'll code this up.

my $parser = Bio::Tools::Phylo::PAML::IO(-format => 'codeml',
				     -file   => $file,
				     -dir    => $dir);


# get a Bio::Tools::Phylo::PAML::Result::Codeml result
# which ISA Bio::Tools::Phylo::PAML::Result::ResultI object.
# (unless qualitative all the PAML results can be represented
#  by the same basic methods? - this would be analagous to
#  the GenericHSP object holding FASTA and BLAST HSPs)

my $codemlresult = $parser->next_result();

The methods you describe for the codeml object are great in my mind.


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