An example of extracting UTR coordinates from an Ensembl transcript

gistfile1.pl

use Bio::EnsEMBL::Registry;
Bio::EnsEMBL::Registry->load_registry_from_db(
-HOST => 'ensembldb.ensembl.org',
-PORT => 5306,
-USER => 'anonymous',
-VERBOSE => 0,
-DB_VERSION => 64
);

my $dba = Bio::EnsEMBL::Registry->get_DBAdaptor('human', 'core');
my $ta = $dba->get_TranscriptAdaptor();

my @ids = qw/ENST00000330803 ENST00000366472 ENST00000318244/;
foreach my $id (@ids) {
  my $t = $ta->fetch_by_stable_id($id);
  
  my $strand = $t->strand();
  
  my $seq_start = $t->seq_region_start();
  my $seq_end = $t->seq_region_end();
  my ($five_prime_coordinate) = $t->cdna2genomic($t->cdna_coding_start(), $t->cdna_coding_start());
  my ($three_prime_coordinate) = $t->cdna2genomic($t->cdna_coding_end(), $t->cdna_coding_end());
  
  my ($five_start, $five_end, $three_start, $three_end) =( (0)x4 );
  
  if($strand == 1) {
    if($seq_start != $five_prime_coordinate->start()) {
      $five_start = $seq_start;
      $five_end = $five_prime_coordinate->start()-1;
    }
    if($seq_end != $three_prime_coordinate->start()) {
      $three_start = $three_prime_coordinate->start()+1;
      $three_end = $seq_end;
    }
  }
  else {
    if($seq_start != $five_prime_coordinate->start()) {
      $five_start = $seq_end;
      $five_end = $five_prime_coordinate->start()+1;
    }
    if($seq_end != $three_prime_coordinate->start()) {
      $three_start = $three_prime_coordinate->start()-1;
      $three_end = $seq_start;
    }
  }
  
  printf(q{5' UTR %d-%d | 3' UTR %d-%d}, $five_start, $five_end, $three_start, $three_end);
  print "\n";
}