Mercurial > repos > genouest > feelnc2asko

comparison feelnc2asko.pl @ 0:e323c49b8bcc draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/genouest/galaxy-tools/tree/master/tools/feelnc2asko commit 92849224db1963d090fbb25d410cc659a5449241
author genouest
date Thu, 12 Apr 2018 06:05:23 -0400
parents
children af75f883cab4
comparison
equal deleted inserted replaced
-1:000000000000 0:e323c49b8bcc
1 use strict;
2 use warnings;
3 use Getopt::Long;
4 use Bio::Tools::GFF;
5
6 my ($anngff, $lncgff, $newgff);
7
8 GetOptions("ann=s" => \$anngff, "lnc=s" => \$lncgff, "new=s" => \$newgff);
9
10 my $gffout = Bio::Tools::GFF->new(-fh=> \*STDOUT, -gff_version => 3);
11
12 #1. the standard annotation
13 my $gffin = Bio::Tools::GFF->new(-file => $anngff, -gff_version => 3);
14
15
16 while (my $feature = $gffin->next_feature()) {
17 if ($feature-> primary_tag eq 'mRNA') {
18 my ($gene)=$feature->get_tag_values("gene");
19 $feature->remove_tag("Parent");
20 $feature->add_tag_value("Parent", $gene);
21 $feature->add_tag_value("feelnc_type", "standard");
22 $gffout->write_feature($feature);
23 }
24 if ($feature-> primary_tag eq 'gene') {
25 my ($name)=$feature->get_tag_values("Name");
26 $feature->remove_tag("ID");
27 $feature->add_tag_value("ID", $name);
28 $feature->add_tag_value("feelnc_type", "standard");
29 $gffout->write_feature($feature);
30 }
31 }
32 $gffin->close();
33
34
35 my %genes=();
36 my %transcripts=();
37
38 #2. The lncRNA gtf
39 my $fncgtf = Bio::Tools::GFF->new( -file => $lncgff, -gff_version => '2' );
40
41 while (my $feat = $fncgtf->next_feature()) {
42 next if ($feat->primary_tag() ne 'exon');
43
44 my $mrna= ($feat->get_tag_values('transcript_id'))[0];
45 my $gene= ($feat->get_tag_values('gene_id'))[0];
46 # print STDERR "str: ", $feat->strand(), "\n";
47 if (exists($genes{$gene})) {
48 if ($genes{$gene}->start() > $feat->start()) {
49 $genes{$gene}->start($feat->start());
50 }
51 if ($genes{$gene}->end() < $feat->end()) {
52 $genes{$gene}->end($feat->end());
53 }
54 }
55 else {
56 my $geneft = Bio::SeqFeature::Generic->new(
57 -start => $feat->start(),
58 -end => $feat->end(),
59 -strand => $feat->strand(),
60 -primary_tag => 'gene',
61 -source_tag => $feat->source_tag(),
62 -seq_id => $feat->seq_id());
63 # $geneft->add_tag_value("feelnc_type", "lncRNA");
64 $genes{$gene}=$geneft;
65 }
66
67 if (exists($transcripts{$mrna})) {
68 if ($transcripts{$mrna}->start() > $feat->start()) {
69 $transcripts{$mrna}->start($feat->start());
70 }
71 if ($transcripts{$mrna}->end() < $feat->end()) {
72 $transcripts{$mrna}->end($feat->end());
73 }
74 }
75 else {
76 my $tr = Bio::SeqFeature::Generic->new(
77 -start => $feat->start(),
78 -end => $feat->end(),
79 -strand => $feat->strand(),
80 -primary_tag => 'mRNA',
81 -source_tag => $feat->source_tag(),
82 -seq_id => $feat->seq_id());
83 $tr->add_tag_value("ID", $mrna);
84 $tr->add_tag_value("Parent",$gene);
85 $tr->add_tag_value("feelnc_type", "lncRNA");
86 $transcripts{$mrna}=$tr;
87 }
88 }
89
90 #3. The new mRNA gtf
91 my $nmgtf = Bio::Tools::GFF->new( -file => $newgff, -gff_version => '2' );
92
93 while (my $feat = $nmgtf->next_feature()) {
94 next if ($feat->primary_tag() ne 'exon');
95
96 my $mrna= ($feat->get_tag_values('transcript_id'))[0];
97 my $gene= ($feat->get_tag_values('gene_id'))[0];
98 # print STDERR "str: ", $feat->strand() , "\n";
99 if (exists($genes{$gene})) {
100 if ($genes{$gene}->start() > $feat->start()) {
101 $genes{$gene}->start($feat->start());
102 }
103 if ($genes{$gene}->end() < $feat->end()) {
104 $genes{$gene}->end($feat->end());
105 }
106 }
107 else {
108 my $geneft = Bio::SeqFeature::Generic->new(
109 -start => $feat->start(),
110 -end => $feat->end(),
111 -strand => $feat->strand(),
112 -primary_tag => 'gene',
113 -source_tag => $feat->source_tag(),
114 -seq_id => $feat->seq_id());
115 $geneft->add_tag_value("ID", $gene);
116 # $geneft->add_tag_value("feelnc_type", "new");
117 $genes{$gene}=$geneft;
118 }
119
120 if (exists($transcripts{$mrna})) {
121 if ($transcripts{$mrna}->start() > $feat->start()) {
122 $transcripts{$mrna}->start($feat->start());
123 }
124 if ($transcripts{$mrna}->end() < $feat->end()) {
125 $transcripts{$mrna}->end($feat->end());
126 }
127 }
128 else {
129 my $tr = Bio::SeqFeature::Generic->new(
130 -start => $feat->start(),
131 -end => $feat->end(),
132 -strand => $feat->strand(),
133 -primary_tag => 'mRNA',
134 -source_tag => $feat->source_tag(),
135 -seq_id => $feat->seq_id());
136 $tr->add_tag_value("ID", $mrna);
137 $tr->add_tag_value("Parent",$gene);
138 $tr->add_tag_value("feelnc_type", "new");
139 $transcripts{$mrna}=$tr;
140 }
141 }
142
143 foreach my $mrna (keys %transcripts) {
144 my ($parent)=$transcripts{$mrna}->get_tag_values("Parent");
145 $gffout->write_feature($genes{$parent});
146 $gffout->write_feature($transcripts{$mrna});
147 }