[Bioperl-l] bioperl dna pairwise local alignment

[Mike Pheasant]

Hi Ewan,

Thanks, I want to align pairs of 50mers. Actually I'd like the best scoring
pairs aligning thousands and thousands of them, so performance will be an
issue. I havent seen the wise2 package, am looking at emboss water.c, on the
surface of it, seems well written and works for dna & protein.

Is Wise2 better? and if so, in what way?

I have got the Bio::Factory::EMBOSS working, thanks for the tip, but I want
a solution that doesnt use `backticks` and pass parameters via file io. By
the way there are deprecated warnings and a prototype ($input) error using
perl 5.8.0. If I fix that I'll send it in.

I think I'm going to try linking it into PostgresQL and/or Mysql as
functions, as I'd like to do joins on high scoring rows. If that works,
might also try XS into perl as well. If I get the perl one working I'll
forward it on.

Anyone know of anyone linking bioinformatic functions directly into the
databases as SQL functions?

Cheers,

Mike.

> -----Original Message-----
> From: Ewan Birney [mailto:birney@ebi.ac.uk]
> Sent: Wednesday, 23 October 2002 5:50 PM
> To: Elia Stupka
> Cc: Mike Pheasant; bioperl-l@bioperl.org
> Subject: Re: [Bioperl-l] bioperl dna pairwise local alignment
>
>
>
>
> On Wed, 23 Oct 2002, Elia Stupka wrote:
>
> > > Has anyone linked the water.c code from EMBOSS into perl ?
> > > Are there any other perl modules anyone knows about which do this?
> >
> > No, but you can run EMBOSS within bioperl, see the synopsys of
> the EMBOSS
> > module, Bio::Factory::EMBOSS, which has an example using water...
>
> Unless you are doing something pretty impressive (like aligning 50mers)
> then the systems overhead should be small vs the alignment code.
>
>
> BTW - you can force pSW to do the alignments by "pretending" that the
> dna sequences are protein and making a "protein" matrix which looks like
> DNA, but I suspect that you will want things like gap-end penalty and
> semi-local alignments.
>
>
> If you feel up to it - take a look at Wise2 package and the program dnal -
> you then need to build XS bindings into that function - get an idea from
> things like gwrap.xs in the wise2/src/models directory. not for the faint
> hearted of course.
>
>
>
>
> >
> > Elia
> >
> > ********************************
> > * http://www.fugu-sg.org/~elia *
> > * tel:    +65 6874 1467        *
> > * mobile: +65 9030 7613        *
> > * fax:    +65 6779 1117        *
> > ********************************
> >
> >
> > _______________________________________________
> > Bioperl-l mailing list
> > Bioperl-l@bioperl.org
> > http://bioperl.org/mailman/listinfo/bioperl-l
> >
>