|
0
|
1 # $Id: Exon.pm,v 1.8 2002/10/22 07:38:41 lapp Exp $
|
|
|
2 #
|
|
|
3 # BioPerl module for Bio::SeqFeature::Gene::Exon
|
|
|
4 #
|
|
|
5 # Cared for by Hilmar Lapp <hlapp@gmx.net>
|
|
|
6 #
|
|
|
7 # Copyright Hilmar Lapp
|
|
|
8 #
|
|
|
9 # You may distribute this module under the same terms as perl itself
|
|
|
10
|
|
|
11 # POD documentation - main docs before the code
|
|
|
12
|
|
|
13 =head1 NAME
|
|
|
14
|
|
|
15 Bio::SeqFeature::Gene::Exon - a feature representing an exon
|
|
|
16
|
|
|
17 =head1 SYNOPSIS
|
|
|
18
|
|
|
19 # obtain an exon instance $exon somehow
|
|
|
20 print "exon from ", $exon->start(), " to ", $exon->end(),
|
|
|
21 " on seq ", $exon->seq_id(), ", strand ", $exon->strand(),
|
|
|
22 ", encodes the peptide sequence ",
|
|
|
23 $exon->cds()->translate()->seq(), "\n";
|
|
|
24
|
|
|
25 =head1 DESCRIPTION
|
|
|
26
|
|
|
27 This module implements a feature representing an exon by implementing the
|
|
|
28 Bio::SeqFeature::Gene::ExonI interface.
|
|
|
29
|
|
|
30 Apart from that, this class also implements Bio::SeqFeatureI by inheriting
|
|
|
31 off Bio::SeqFeature::Generic.
|
|
|
32
|
|
|
33 =head1 FEEDBACK
|
|
|
34
|
|
|
35 =head2 Mailing Lists
|
|
|
36
|
|
|
37 User feedback is an integral part of the evolution of this
|
|
|
38 and other Bioperl modules. Send your comments and suggestions preferably
|
|
|
39 to one of the Bioperl mailing lists.
|
|
|
40 Your participation is much appreciated.
|
|
|
41
|
|
|
42 bioperl-l@bioperl.org - General discussion
|
|
|
43 http://bio.perl.org/MailList.html - About the mailing lists
|
|
|
44
|
|
|
45 =head2 Reporting Bugs
|
|
|
46
|
|
|
47 Report bugs to the Bioperl bug tracking system to help us keep track
|
|
|
48 the bugs and their resolution.
|
|
|
49 Bug reports can be submitted via email or the web:
|
|
|
50
|
|
|
51 bioperl-bugs@bio.perl.org
|
|
|
52 http://bugzilla.bioperl.org/
|
|
|
53
|
|
|
54 =head1 AUTHOR - Hilmar Lapp
|
|
|
55
|
|
|
56 Email hlapp@gmx.net
|
|
|
57
|
|
|
58 Describe contact details here
|
|
|
59
|
|
|
60 =head1 APPENDIX
|
|
|
61
|
|
|
62 The rest of the documentation details each of the object methods.
|
|
|
63 Internal methods are usually preceded with a _
|
|
|
64
|
|
|
65 =cut
|
|
|
66
|
|
|
67
|
|
|
68 # Let the code begin...
|
|
|
69
|
|
|
70
|
|
|
71 package Bio::SeqFeature::Gene::Exon;
|
|
|
72 use vars qw(@ISA);
|
|
|
73 use strict;
|
|
|
74
|
|
|
75 use Bio::SeqFeature::Generic;
|
|
|
76 use Bio::SeqFeature::Gene::ExonI;
|
|
|
77
|
|
|
78 @ISA = qw(Bio::SeqFeature::Generic Bio::SeqFeature::Gene::ExonI);
|
|
|
79
|
|
|
80 #
|
|
|
81 # A list of allowed exon types. See primary_tag().
|
|
|
82 #
|
|
|
83 my @valid_exon_types = ('initial', 'internal', 'terminal');
|
|
|
84
|
|
|
85 sub new {
|
|
|
86 my ($caller, @args) = @_;
|
|
|
87 my $self = $caller->SUPER::new(@args);
|
|
|
88 my ($primary) =
|
|
|
89 $self->_rearrange([qw(PRIMARY)],@args);
|
|
|
90
|
|
|
91 $primary = 'exon' unless $primary;
|
|
|
92 $self->primary_tag($primary); # this will also set is_coding()
|
|
|
93 $self->strand(0) if(! defined($self->strand()));
|
|
|
94 return $self;
|
|
|
95 }
|
|
|
96
|
|
|
97
|
|
|
98 =head2 is_coding
|
|
|
99
|
|
|
100 Title : is_coding
|
|
|
101 Usage : if($exon->is_coding()) {
|
|
|
102 # do something
|
|
|
103 }
|
|
|
104 if($is_utr) {
|
|
|
105 $exon->is_coding(0);
|
|
|
106 }
|
|
|
107 Function: Get/set whether or not the exon codes for amino acid.
|
|
|
108 Returns : TRUE if the object represents a feature translated into protein,
|
|
|
109 and FALSE otherwise.
|
|
|
110 Args : A boolean value on set.
|
|
|
111
|
|
|
112
|
|
|
113 =cut
|
|
|
114
|
|
|
115 sub is_coding {
|
|
|
116 my ($self,$val) = @_;
|
|
|
117
|
|
|
118 if(defined($val)) {
|
|
|
119 $self->{'_iscoding'} = $val;
|
|
|
120 }
|
|
|
121 return $self->{'_iscoding'};
|
|
|
122 }
|
|
|
123
|
|
|
124 =head2 primary_tag
|
|
|
125
|
|
|
126 Title : primary_tag
|
|
|
127 Usage : $tag = $feat->primary_tag()
|
|
|
128 $feat->primary_tag('exon')
|
|
|
129 Function: Get/set the primary tag for the exon feature.
|
|
|
130
|
|
|
131 This method is overridden here in order to allow only for
|
|
|
132 tag values following a certain convention. For consistency reasons,
|
|
|
133 the tag value must either contain the string 'exon' or the string
|
|
|
134 'utr' (both case-insensitive). In the case of 'exon', a string
|
|
|
135 describing the type of exon may be appended or prefixed. Presently,
|
|
|
136 the following types are allowed: initial, internal, and terminal
|
|
|
137 (all case-insensitive).
|
|
|
138
|
|
|
139 If the supplied tag value matches 'utr' (case-insensitive),
|
|
|
140 is_coding() will automatically be set to FALSE, and to TRUE
|
|
|
141 otherwise.
|
|
|
142
|
|
|
143 Returns : A string.
|
|
|
144 Args : A string on set.
|
|
|
145
|
|
|
146
|
|
|
147 =cut
|
|
|
148
|
|
|
149 # sub primary_tag {
|
|
|
150 # my ($self,$value) = @_;
|
|
|
151
|
|
|
152 # if(defined($value)) {
|
|
|
153 # if((lc($value) =~ /utr/i) || (lc($value) eq "exon") ||
|
|
|
154 # ((lc($value) =~ /exon/i) &&
|
|
|
155 # (grep { $value =~ /$_/i; } @valid_exon_types))) {
|
|
|
156 # $self->is_coding($value =~ /utr/i ? 0 : 1);
|
|
|
157 # } else {
|
|
|
158 # $self->throw("primary tag $value is invalid for object of class ".
|
|
|
159 # ref($self));
|
|
|
160 # }
|
|
|
161 # }
|
|
|
162 # return $self->SUPER::primary_tag($value);
|
|
|
163 # }
|
|
|
164
|
|
|
165 =head2 location
|
|
|
166
|
|
|
167 Title : location
|
|
|
168 Usage : my $location = $exon->location()
|
|
|
169 Function: Returns a location object suitable for identifying the location
|
|
|
170 of the exon on the sequence or parent feature.
|
|
|
171
|
|
|
172 This method is overridden here to restrict allowed location types
|
|
|
173 to non-compound locations.
|
|
|
174
|
|
|
175 Returns : Bio::LocationI object
|
|
|
176 Args : none
|
|
|
177
|
|
|
178
|
|
|
179 =cut
|
|
|
180
|
|
|
181 sub location {
|
|
|
182 my ($self,$value) = @_;
|
|
|
183
|
|
|
184 if(defined($value) && $value->isa('Bio::Location::SplitLocationI')) {
|
|
|
185 $self->throw("split or compound location is not allowed ".
|
|
|
186 "for an object of type " . ref($self));
|
|
|
187 }
|
|
|
188 return $self->SUPER::location($value);
|
|
|
189 }
|
|
|
190
|
|
|
191 =head2 cds
|
|
|
192
|
|
|
193 Title : cds()
|
|
|
194 Usage : $cds = $exon->cds();
|
|
|
195 Function: Get the coding sequence of the exon as a sequence object.
|
|
|
196
|
|
|
197 The sequence of the returned object is prefixed by Ns (lower case)
|
|
|
198 if the frame of the exon is defined and different from zero. The
|
|
|
199 result is that the first base starts a codon (frame 0).
|
|
|
200
|
|
|
201 This implementation returns undef if the particular exon is
|
|
|
202 not translated to protein, i.e., is_coding() returns FALSE. Undef
|
|
|
203 will also be returned if no sequence is attached to this exon
|
|
|
204 feature.
|
|
|
205
|
|
|
206 Returns : A Bio::PrimarySeqI implementing object.
|
|
|
207 Args :
|
|
|
208
|
|
|
209
|
|
|
210 =cut
|
|
|
211
|
|
|
212 sub cds {
|
|
|
213 my ($self) = @_;
|
|
|
214
|
|
|
215 # UTR is not translated
|
|
|
216 return undef if(! $self->is_coding());
|
|
|
217
|
|
|
218 my $seq = $self->seq();
|
|
|
219 if(defined($seq) && defined($self->frame()) && ($self->frame() != 0)) {
|
|
|
220 my $prefix = "n" x $self->frame();
|
|
|
221 $seq->seq($prefix . $seq->seq());
|
|
|
222 }
|
|
|
223 return $seq;
|
|
|
224 }
|
|
|
225
|
|
|
226 1;
|
