[Bioperl-l] Phylip error

[Jason Stajich]

Does the protdist test work fine for you in bioperl-run?

Are you using phylip 3.5 or 3.6?  You need to twiddle something to use 3.6
bash% export PHYLIPVERSION = 3.6

or in your script:
$ENV{PHYLIPVERSION} = 3.6;
use Bio::Tools::Run::Phylo::Phylip::ProtDist;

If you are just doing kimura 2P you can avoid ProtDist completely with new
code in Bio::Align::ProteinStatistics.

-jason
On Wed, 2 Jun 2004, Jonathan Manning wrote:

> Hi to all,
>
> I'm trying to incorporate phylogenetic analysis into a script I'm
> writing. I'm attempting to feed in alignments (SimpleAlign objects)
> derived from an earlier subroutine, and stored in a global hash (the
> reason for the hash being that I wanted to reference by a blast query
> sequence). My code:
>
> sub phylip{
>
> 	foreach my $key (keys %alignments){
> 		my ($alnmnt) = $alignments{$key};
> 		my @params = ('MODEL' => 'PAM');
> 		my $protdist = Bio::Tools::Run::Phylo::Phylip::ProtDist->new(@params);
>   		my $matrix  = $protdist->run($alnmnt);
> 		@params = ('type'=>'NJ','outgroup'=>2,'lowtri'=>1,
>              'upptri'=>1,'subrep'=>1);
>
>   		my $neighbor =
> Bio::Tools::Run::Phylo::Phylip::Neighbor->new(@params);
> 		my ($tree) = $neighbor_factory->run($matrix);
> 	}
>
> }
>
>
>
> It seems to have a problem at the protdist stage, resulting in the
> following error:
>
>
>
> ------------- EXCEPTION  -------------
> MSG: protdist did not create matrix correctly (/tmp/q4lIZVPNvW/outfile)
> STACK Bio::Tools::Run::Phylo::Phylip::ProtDist::_run
> /usr/lib/perl5/site_perl/5.8.3/Bio/Tools/Run/Phylo/Phylip/ProtDist.pm:407
> STACK Bio::Tools::Run::Phylo::Phylip::ProtDist::run
> /usr/lib/perl5/site_perl/5.8.3/Bio/Tools/Run/Phylo/Phylip/ProtDist.pm:364
> STACK main::phylip ./GeneInfo5.pl:1648
> STACK toplevel ./GeneInfo5.pl:563
>
> --------------------------------------
>
>
> >From the archives, this would seem to be a problem with writing the
> matrix to file, but I don't know why. The permissions on /tmp are fine.
>
> Does anyone have any ideas?
>
> Thanks in advance,
>
> Jon
>
> _______________________________________________
> Bioperl-l mailing list
> Bioperl-l at portal.open-bio.org
> http://portal.open-bio.org/mailman/listinfo/bioperl-l
>

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