use Bio::Tools::Geneid;
  my $gid = Bio::Tools::Geneid(-file => "geneid.out");

  while (my $gene = $gid->next_prediction)
  {
    my @transcripts = $gene->transcripts;
      foreach my $t (@transcripts)
      {
        my @exons = $t->exons;
        foreach my $e (@exons)
        {
          printf("Exon %d..%d\n", $e->start, $e->end);
        }
      }
  }

    my($class, @args) = @_;
    my $self = $class->SUPER::new(@args);
    $self->_initialize_io(@args);
    return $self;

    my ($self) = @_;

    my ($gene, $transcript, $current_gene_id);
    my $transcript_score = 0;

    my ($gene_id, $exon_type, $exon_start, $exon_end, $exon_score,
        $exon_strand, $start_phase, $end_phase, $start_sig_score,
        $end_sig_score, $coding_pot_score, $homol_score);

    while (defined($_ = $self->_readline))
    {
        $self->debug($_) if ($self->verbose > 0);

        s/^\s+//;
        s/\s+$//;

        # We have a choice of geneid, gff or XML formats. The native
        # geneid format has more information than gff. However, we
        # then need to perform the hack of extracting the sequence ID
        # from the header of the embedded Fasta file which comes after
        # the exon data, as it is not stored elsewhere. Ack.
        if (/^>(\S+)\|GeneId/)
        {
            my $target_id = $1;
            $self->_target_id($target_id) unless defined $self->_target_id;
            next;
        }

        next unless (/(Single|First|Internal|Terminal)/);

        my @fields = split(/\s+/, $_);

        # Grab gene_id from eol first as there are issues with
        # inconsistent whitespace in the AA coords field
        $gene_id = pop @fields;

        ($exon_type, $exon_start, $exon_end, $exon_score,
         $exon_strand, $start_phase, $end_phase, $start_sig_score,
         $end_sig_score, $coding_pot_score, $homol_score) = @fields[0..10];

        if (! defined $current_gene_id)
        {
            # Starting the requested prediction
            $current_gene_id = $gene_id;
            $transcript_score = $exon_score;

            $gene = Bio::SeqFeature::Gene::GeneStructure->new(-source =>
                                                              $SOURCE_TAG);
            $transcript = Bio::SeqFeature::Gene::Transcript->new(-source =>
                                                                 $SOURCE_TAG);

            $self->_add_exon($gene, $transcript, $exon_type, $exon_start, $exon_end, $exon_score,
                             $exon_strand, $start_phase, $end_phase, $start_sig_score,
                             $end_sig_score, $coding_pot_score, $homol_score);
        }
        elsif ($gene_id eq $current_gene_id)
        {
            # Still in requested prediction
            $transcript_score += $exon_score;

            $self->_add_exon($gene, $transcript, $exon_type, $exon_start, $exon_end, $exon_score,
                             $exon_strand, $start_phase, $end_phase, $start_sig_score,
                             $end_sig_score, $coding_pot_score, $homol_score);
        }
        else
        {
            # Found following prediction
            $self->_pushback($_);
            last;
        }
    }

    if (defined $gene)
    {
        $transcript->seq_id($self->_target_id);
        $transcript->score($transcript_score);
        $gene->add_transcript($transcript);
        $gene->seq_id($self->_target_id);

        foreach my $exon ($gene->exons)
        {
            $exon->seq_id($self->_target_id);
        }

        $self->_set_strand($gene);
    }

    return $gene;

    my ($self, $gene, $transcript, $exon_type, $exon_start, $exon_end,
        $exon_score, $exon_strand, $start_phase, $end_phase, $start_sig_score,
        $end_sig_score, $coding_pot_score, $homol_score) = @_;

    $exon_type =~ s/First/Initial/;

    my $strand = $exon_strand eq '+' ? 1 : -1;

    my $exon = Bio::SeqFeature::Gene::Exon->new(-source => $SOURCE_TAG,
                                                -start  => $exon_start,
                                                -end    => $exon_end,
                                                -strand => $strand,
                                                -score  => $exon_score);
    $exon->is_coding(1);
    $exon->add_tag_value("Type", $exon_type);
    $exon->add_tag_value('phase', $start_phase);
    $exon->add_tag_value('end_phase', $end_phase);
    $exon->add_tag_value('start_signal_score', $start_sig_score);
    $exon->add_tag_value('end_signal_score', $end_sig_score);
    $exon->add_tag_value('coding_potential_score', $coding_pot_score);
    $exon->add_tag_value('homology_score', $homol_score);

    $transcript->strand($strand) unless $transcript->strand != 0;
    $transcript->add_exon($exon, $exon_type);

    my ($self, $gene) = @_;

    my $fwd = 0;
    my $rev = 0;

    my @exons = $gene->exons;
    foreach my $exon (@exons)
    {
        my $strand = $exon->strand;

        if ($strand == 1)
        {
            $fwd++;
        }
        elsif ($strand == -1)
        {
            $rev++;
        }
    }

    if ($#exons == $fwd)
    {
        $gene->strand(1);
    }
    elsif ($#exons == $rev)
    {
        $gene->strand(-1);
    }
    else
    {
        $gene->strand(0);
    }

    return $gene;

    my ($self,$val) = @_;
    if ($val)
    {
        $self->{'_target_id'} = $val;
    }

    return $self->{'_target_id'};

  bioperl-l@bioperl.org                  - General discussion
  http://bioperl.org/wiki/Mailing_lists  - About the mailing lists

  http://bugzilla.open-bio.org/

 Email: kdj@sanger.ac.uk