0
|
1 =head1 LICENSE
|
|
2
|
|
3 Copyright (c) 1999-2012 The European Bioinformatics Institute and
|
|
4 Genome Research Limited. All rights reserved.
|
|
5
|
|
6 This software is distributed under a modified Apache license.
|
|
7 For license details, please see
|
|
8
|
|
9 http://www.ensembl.org/info/about/code_licence.html
|
|
10
|
|
11 =head1 CONTACT
|
|
12
|
|
13 Please email comments or questions to the public Ensembl
|
|
14 developers list at <dev@ensembl.org>.
|
|
15
|
|
16 Questions may also be sent to the Ensembl help desk at
|
|
17 <helpdesk@ensembl.org>.
|
|
18
|
|
19 =cut
|
|
20
|
|
21 =head1 NAME
|
|
22
|
|
23 =head1 SYNOPSIS
|
|
24
|
|
25 =head1 DESCRIPTION
|
|
26
|
|
27 =head1 METHODS
|
|
28
|
|
29 =cut
|
|
30
|
|
31 package Bio::EnsEMBL::Utils::VegaCuration::Gene;
|
|
32
|
|
33 use strict;
|
|
34 use warnings;
|
|
35 use vars qw(@ISA);
|
|
36
|
|
37 use Bio::EnsEMBL::Utils::ConversionSupport;
|
|
38
|
|
39 @ISA = qw(Bio::EnsEMBL::Utils::ConversionSupport);
|
|
40
|
|
41
|
|
42 =head2 find_gaps
|
|
43
|
|
44 Args : arrayref of B::E::Transcripts
|
|
45 Example : my $gaps = find_gaps($all_transcripts)
|
|
46 Description: identifies regions of a gene that are not covered by any transcript
|
|
47 Returntype : int
|
|
48 Exceptions : none
|
|
49 Caller : internal
|
|
50
|
|
51 =cut
|
|
52
|
|
53 sub find_gaps {
|
|
54 my $self = shift;
|
|
55 my ($all_transcripts) = @_;
|
|
56 my $gaps = 0;
|
|
57 my @sorted_transcripts = sort {$a->start <=> $b->start || $b->end <=> $a->end} @{$all_transcripts};
|
|
58 if ( my $first_transcript = shift @sorted_transcripts ) {
|
|
59 my $pos = $first_transcript->end;
|
|
60 foreach my $transcript (@sorted_transcripts) {
|
|
61 next if ($transcript->end < $pos );
|
|
62 if ($transcript->start < $pos && $transcript->end > $pos ) {
|
|
63 $pos = $transcript->end;
|
|
64 next;
|
|
65 }
|
|
66 elsif ($transcript->end > $pos) {
|
|
67 $gaps++;
|
|
68 $pos = $transcript->end;
|
|
69 }
|
|
70 }
|
|
71 }
|
|
72 return $gaps;
|
|
73 }
|