[Bioperl-l] Error when parsing MEME file

Jason Stajich jason at cgt.duhs.duke.edu
Thu May 6 08:29:04 EDT 2004


Something was still wrong with your fix stefan - I committed something
which should work.  Do a CVS update Amonida and you should have thefix.

As for whether or not Seqio can parse long IDs, it just takes whatever is
after the > until the 1st space.  We don't try and interpret the ID and
get accession out of the ncbi id lines by default - although there have
been proposals on the list for doing this, I can't really remember if
anyone solved it.  I just end up writing my own id parser if I need a
particular thing out...  We solve it for the HitI object when parsing
blast reports and setting accessions.

-jason

On Thu, 6 May 2004, skirov wrote:

> Amonida,
> Wow, you have an amazing amount of ids... I am sure it will be much better if
> you can use a simple ID line like
> >X61109
> someseq...
> Of course I could change the parser... but I am not sure it is worth it.
> Jason, forgive my ignorance, but current SeqIO will not parse fasta ID line
> like this one:
> >ref|NW_047562.1|Rn1_2084
> I guess it will be best if meme parser is in sync with SeqIO with respect to
> IDs.
> Stefan
>
> >===== Original Message From Amonida Zadissa <amonida at sanger.otago.ac.nz>
> =====
> >Stefan and Jason,
> >
> >Thanks for your replies. I was using bioperl-1.4 but now I have
> >bioperl-live installed. Nevertheless, the problem seems to be there
> >but I get new warnings complainging about both strands now.
> >
> >Argument "-" isn't numeric in addition (+) at
> >/opt/share/perl/5.8.3/Bio/LocatableSeq.pm line 167, <GEN0> line 699.
> >Argument "+" isn't numeric in addition (+) at
> >/opt/share/perl/5.8.3/Bio/Matrix/PSM/InstanceSite.pm line 106, <GEN0>
> >line 700.
> >
> >Could this depend on my installation (which by the way went fine)?
> >
> >I'm attaching the MEME test file I have used.
> >
> >Thanks for any help,
> >Amonida
> >
> >On Wed, May 05, 2004 at 02:51:12PM -0400, Stefan Kirov wrote:
> >> Fixed now, check it from bioperl-live
> >> Stefan
> >>
> >> Amonida Zadissa wrote:
> >>
> >> >Hi everyone,
> >> >
> >> >I'm trying to parse a MEME file using the Bio::Matrix::PSM
> >> >modules but I'm experiencing problems and was hoping someone on
> >> >the list could help me.
> >> >
> >> >The script is attached below. I get a warning saying the
> >> >following for each "problematic" instance:
> >> >
> >> >Argument "-" isn't numeric in addition (+) at
> >> >/opt/share/perl/5.8.3/Bio/Matrix/PSM/InstanceSite.pm line 106, <GEN0> line
> >> >1164.
> >> >
> >> >Whenever the strand sign is negative the start position is
> >> >assigned the sign and the score is assigned the start position,
> >> >so everything seems to shift one column. This happens only
> >> >if the strand is negative. Note that '$instance' becomes a
> >> >Bio::Matrix::PSM::InstanceSite object.
> >> >
> >> >Is this a parsing error in "InstanceSite.pm" or is my script
> >> >somehow wrong?
> >> >
> >> >Would appreciate any help and by the way thanks for providing this
> >> >forum!
> >> >
> >> >Cheers,
> >> >Amonida
> >> >
> >> >********************************************************************
> >> >#!/usr/bin/perl -w
> >> >
> >> >use strict;
> >> >use warnings;
> >> >use Bio::Matrix::PSM::IO;
> >> >
> >> >my $psmIO = new Bio::Matrix::PSM::IO(	-file   => 'meme.file',
> >> >					-format => 'meme');
> >> >
> >> >while (my $psm = $psmIO -> next_psm) {
> >> >   my $instances = $psm -> instances;
> >> >   foreach my $instance (@{ $instances }) {
> >> >   my $start = $instance -> start;
> >> >   my $score = $instance -> score;
> >> >   }
> >> >}
> >> >
> >> >********************************************************************
> >> >
> >> >
> >> >
> >
> >--
> >Amonida Zadissa
> >Biochemistry Department
> >Otago University
> >Dunedin, New Zealand
>
> _______________________________________________
> 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


More information about the Bioperl-l mailing list