protein_sequence

#!/usr/bin/perl

use strict;
use warnings;
use Getopt::Long;

use FindBin;
use lib "$FindBin::RealBin/../lib";
use GAL::Annotation;

#-----------------------------------------------------------------------------
#----------------------------------- MAIN ------------------------------------
#-----------------------------------------------------------------------------

my $usage = "

Synopsis:

gal_protein_sequence feature.gff3 sequence.fasta

Description:

Pint the protein sequence of protein coding genes.

";

my ($help, $clean);
my $opt_success = GetOptions('help'          => \$help,
			     'clean'         => \$clean,
			      );

if (! $opt_success) {
    print STDERR join ' : ', ('FATAL',
                              'command_line_parse_error',
                              'Use gal_protein_genes --help to see correct usage');
}

if ($help || !@ARGV) {
 print $usage;
 exit(0);
}

my ($gff3, $fasta) = @ARGV;

die $usage unless $fasta && $gff3;

my $annotation = GAL::Annotation->new($gff3,
				      $fasta);

my $features = $annotation->features;

my $genes = $features->search({type => 'gene'});
GENE:
while (my $gene = $genes->next) {
  next GENE unless $gene->is_coding;
  my $mrnas = $gene->mRNAs;
 MRNA:
  while (my $mrna = $mrnas->next) {
    my $mrna_id = $mrna->feature_id;
    my $prot_seq = $annotation->wrap_text($mrna->protein_seq);
    print ">$mrna_id (protein translation)\n";
    print $prot_seq;
    print "\n";
  }
}