[Bioperl-l] Re: Substitution matrix format for Bio::Tools::pSW (Wise2)

Jason Stajich jason@cgt.mc.duke.edu
Fri, 8 Feb 2002 12:23:24 -0500 (EST)


I have written an emboss alignment parser so you don't have to use bl2seq
to get pairwise alignments We have emboss application capabilities in the
dev release - this is going to be the way of the future I think.

You can do the following:

#!/usr/bin/perl -w
use strict;
use Bio::Factory::EMBOSS;
use Bio::AlignIO;

my $factory = new Bio::Factory::EMBOSS();
my $water = $factory->program('water');

$water->run({-sequencea   => 'a.fa',
			-seqall      => 'b.fa',
			-datafile    => 'EBLOSUM50',
			-outfile     => 'a_b.emboss',
		    });
if( -e "a_b.emboss" ) {
 my $alignin = new Bio::AlignIO(-format => 'emboss',
			       -file   => 'a_b.emboss');

 my $aln = $alignin->next_aln();

 my $alignout = new Bio::AlignIO(-format => 'clustalw',
				-fh => \*STDOUT);
 $alignout->write_aln($aln);
} else {
	print STDERR "unable to run the emboss program\n";
}

We will eventually want to add some wrapper methods so you won't have to
explicitly dump your sequence files out or instantiate an AlignIO reader.

We're also working with Martin Senger at EBI and Catherine Letondal at
Pasteur to interface with the PISE Web and openBSA CORBA interfaces for
executing programs so that you won't necessarily have to install EMBOSS on
your machine if you don't want to and can execute these analyses remotely
(or on your own farm).

I feel that this is a better way for us to go than bioperl-ext and it
allows us to integrate with other open source package.  Using the emboss
pkg also allows people to get SW DNA alignments which have been past
requests.

-jason
On Fri, 8 Feb 2002, Michal Kurowski wrote:

> Ewan Birney [birney@ebi.ac.uk] wrote:
>
> >
> > how many people use pSW in anger?
> >
>
> Sorry Ewan, but there are some people who do find it useful.
> I've heard Perl6 will have no XS, so someone could rewrite it for Parrot
> ;-).
>
> More seriously, pSW is able to return nice SimpleAlign objects and
> bl2seq for example can be really tricky when it comes to it.
>
>
> Cheers,
>
>

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