|
0
|
1 # Parser module for Coil Bio::Tools::Coil
|
|
|
2 #
|
|
|
3 # Based on the EnsEMBL module Bio::EnsEMBL::Pipeline::Runnable::Protein::Coil
|
|
|
4 # originally written by Marc Sohrmann (ms2@sanger.ac.uk)
|
|
|
5 # Written in BioPipe by Balamurugan Kumarasamy <savikalpa@fugu-sg.org>
|
|
|
6 # Cared for by the Fugu Informatics team (fuguteam@fugu-sg.org)
|
|
|
7
|
|
|
8 # You may distribute this module under the same terms as perl itself
|
|
|
9 #
|
|
|
10 # POD documentation - main docs before the code
|
|
|
11
|
|
|
12 =head1 NAME
|
|
|
13
|
|
|
14 Bio::Tools::Coil
|
|
|
15
|
|
|
16 =head1 SYNOPSIS
|
|
|
17
|
|
|
18 use Bio::Tools::Coil
|
|
|
19 my $parser = new Bio::Tools::Coil();
|
|
|
20 while( my $sp_feat = $parser->next_result($file) ) {
|
|
|
21 #do something
|
|
|
22 #eg
|
|
|
23 push @sp_feat, $sp_feat;
|
|
|
24 }
|
|
|
25
|
|
|
26 =head1 DESCRIPTION
|
|
|
27
|
|
|
28 Parser for Coil output
|
|
|
29
|
|
|
30 =head1 FEEDBACK
|
|
|
31
|
|
|
32 =head2 Mailing Lists
|
|
|
33
|
|
|
34 User feedback is an integral part of the evolution of this and other
|
|
|
35 Bioperl modules. Send your comments and suggestions preferably to
|
|
|
36 the Bioperl mailing list. Your participation is much appreciated.
|
|
|
37
|
|
|
38 bioperl-l@bioperl.org - General discussion
|
|
|
39 http://bioperl.org/MailList.shtml - About the mailing lists
|
|
|
40
|
|
|
41 =head2 Reporting Bugs
|
|
|
42
|
|
|
43 Report bugs to the Bioperl bug tracking system to help us keep track
|
|
|
44 of the bugs and their resolution. Bug reports can be submitted via
|
|
|
45 email or the web:
|
|
|
46
|
|
|
47 bioperl-bugs@bio.perl.org
|
|
|
48 http://bugzilla.bioperl.org/
|
|
|
49
|
|
|
50 =head1 AUTHOR
|
|
|
51
|
|
|
52 Based on the EnsEMBL module Bio::EnsEMBL::Pipeline::Runnable::Protein::Coil
|
|
|
53 originally written by Marc Sohrmann (ms2@sanger.ac.uk)
|
|
|
54 Written in BioPipe by Balamurugan Kumarasamy <savikalpa@fugu-sg.org>
|
|
|
55 Cared for by the Fugu Informatics team (fuguteam@fugu-sg.org)
|
|
|
56
|
|
|
57 =head1 APPENDIX
|
|
|
58
|
|
|
59 The rest of the documentation details each of the object methods.
|
|
|
60 Internal methods are usually preceded with a _
|
|
|
61
|
|
|
62
|
|
|
63 =cut
|
|
|
64
|
|
|
65 package Bio::Tools::Coil;
|
|
|
66 use vars qw(@ISA);
|
|
|
67 use strict;
|
|
|
68
|
|
|
69 use Bio::Root::Root;
|
|
|
70 use Bio::SeqFeature::FeaturePair;
|
|
|
71 use Bio::Root::IO;
|
|
|
72 use Bio::SeqFeature::Generic;
|
|
|
73 @ISA = qw(Bio::Root::Root Bio::Root::IO);
|
|
|
74
|
|
|
75
|
|
|
76
|
|
|
77 sub new {
|
|
|
78 my($class,@args) = @_;
|
|
|
79
|
|
|
80 my $self = $class->SUPER::new(@args);
|
|
|
81 $self->_initialize_io(@args);
|
|
|
82
|
|
|
83 return $self;
|
|
|
84 }
|
|
|
85
|
|
|
86 =head2 parse_results
|
|
|
87
|
|
|
88 Title : parse_results
|
|
|
89 Usage : obj->parse_results
|
|
|
90 Function: Parses the coil output. Automatically called by
|
|
|
91 next_result() if not yet done.
|
|
|
92 Example :
|
|
|
93 Returns :
|
|
|
94
|
|
|
95 =cut
|
|
|
96
|
|
|
97 sub parse_results {
|
|
|
98 my ($self,$resfile) = @_;
|
|
|
99 my $filehandle = $resfile;
|
|
|
100 my %result_hash =_read_fasta($filehandle);#bala no file handle
|
|
|
101 my @ids = keys %result_hash;
|
|
|
102 my @feats;
|
|
|
103 foreach my $id (keys %result_hash){
|
|
|
104 my $pep = reverse ($result_hash{$id});
|
|
|
105 my $count = my $switch = 0;
|
|
|
106 my ($start, $end);
|
|
|
107 while (my $aa = chop $pep) {
|
|
|
108 $count++;
|
|
|
109 if (!$switch && $aa eq "x") {
|
|
|
110 $start = $count;
|
|
|
111 $switch = 1;
|
|
|
112 }
|
|
|
113 elsif ($switch && $aa ne "x") {
|
|
|
114 $end = $count-1;
|
|
|
115 my (%feature);
|
|
|
116 $feature{name} = $id;
|
|
|
117 $feature{start} = $start;
|
|
|
118 $feature{end} = $end;
|
|
|
119 $feature{source} = "Coils";
|
|
|
120 $feature{primary} = 'ncoils';
|
|
|
121 ($feature{program}) = 'ncoils';
|
|
|
122 $feature{logic_name} = 'Coils';
|
|
|
123 my $new_feat = $self->create_feature (\%feature);
|
|
|
124 $self->_add_prediction($new_feat);
|
|
|
125 $switch = 0;
|
|
|
126 }
|
|
|
127 }
|
|
|
128 }
|
|
|
129
|
|
|
130 $self->_predictions_parsed(1);
|
|
|
131
|
|
|
132 }
|
|
|
133 =head2 next_result
|
|
|
134
|
|
|
135 Title : next_result
|
|
|
136 Usage : while($feat = $coil->next_result($file)) {
|
|
|
137 # do something
|
|
|
138 }
|
|
|
139 Function: Returns the next protein feature of the coil output file
|
|
|
140 Returns :
|
|
|
141 Args :
|
|
|
142
|
|
|
143 =cut
|
|
|
144
|
|
|
145 sub next_result{
|
|
|
146
|
|
|
147 my ($self,$resfile) = @_;
|
|
|
148 my $gene;
|
|
|
149
|
|
|
150 $self->parse_results($resfile) unless $self->_predictions_parsed();
|
|
|
151
|
|
|
152 $gene = $self->_result();
|
|
|
153
|
|
|
154 return $gene;
|
|
|
155
|
|
|
156 }
|
|
|
157
|
|
|
158 =head2 _result
|
|
|
159
|
|
|
160 Title : _result
|
|
|
161 Usage : $feat = $obj->_result()
|
|
|
162 Function: internal
|
|
|
163 Example :
|
|
|
164 Returns :
|
|
|
165
|
|
|
166 =cut
|
|
|
167
|
|
|
168 sub _result{
|
|
|
169 my ($self) = @_;
|
|
|
170
|
|
|
171 return undef unless(exists($self->{'_feats'}) && @{$self->{'_feats'}});
|
|
|
172 return shift(@{$self->{'_feats'}});
|
|
|
173
|
|
|
174 }
|
|
|
175
|
|
|
176 =head2 _add_prediction
|
|
|
177
|
|
|
178 Title : _add_prediction()
|
|
|
179 Usage : $obj->_add_prediction($feat)
|
|
|
180 Function: internal
|
|
|
181 Example :
|
|
|
182 Returns :
|
|
|
183
|
|
|
184 =cut
|
|
|
185
|
|
|
186 sub _add_prediction {
|
|
|
187 my ($self, $gene) = @_;
|
|
|
188
|
|
|
189 if(! exists($self->{'_feats'})) {
|
|
|
190 $self->{'_feats'} = [];
|
|
|
191 }
|
|
|
192 push(@{$self->{'_feats'}}, $gene);
|
|
|
193 }
|
|
|
194
|
|
|
195 =head2 _predictions_parsed
|
|
|
196
|
|
|
197 Title : _predictions_parsed
|
|
|
198 Usage : $obj->_predictions_parsed
|
|
|
199 Function: internal
|
|
|
200 Example :
|
|
|
201 Returns : TRUE or FALSE
|
|
|
202
|
|
|
203 =cut
|
|
|
204
|
|
|
205 sub _predictions_parsed {
|
|
|
206 my ($self, $val) = @_;
|
|
|
207
|
|
|
208 $self->{'_preds_parsed'} = $val if $val;
|
|
|
209 if(! exists($self->{'_preds_parsed'})) {
|
|
|
210 $self->{'_preds_parsed'} = 0;
|
|
|
211 }
|
|
|
212 return $self->{'_preds_parsed'};
|
|
|
213 }
|
|
|
214
|
|
|
215
|
|
|
216 =head2 create_feature
|
|
|
217
|
|
|
218 Title : create_feature
|
|
|
219 Usage : obj->create_feature(\%feature)
|
|
|
220 Function: Internal(not to be used directly)
|
|
|
221 Returns :
|
|
|
222 Args :
|
|
|
223
|
|
|
224
|
|
|
225 =cut
|
|
|
226
|
|
|
227 sub create_feature {
|
|
|
228 my ($self, $feat) = @_;
|
|
|
229
|
|
|
230
|
|
|
231 # create feature object
|
|
|
232 my $feature = Bio::SeqFeature::Generic->new(-seq_id => $feat->{name},
|
|
|
233 -start => $feat->{start},
|
|
|
234 -end => $feat->{end},
|
|
|
235 -score => $feat->{score},
|
|
|
236 -source => $feat->{source},
|
|
|
237 -primary => $feat->{primary},
|
|
|
238 -logic_name => $feat->{logic_name},
|
|
|
239 );
|
|
|
240 $feature->add_tag_value('evalue',0);
|
|
|
241 $feature->add_tag_value('percent_id','NULL');
|
|
|
242 $feature->add_tag_value("hid",$feat->{primary});
|
|
|
243
|
|
|
244
|
|
|
245 return $feature;
|
|
|
246
|
|
|
247 }
|
|
|
248
|
|
|
249 =head2 _read_fasta
|
|
|
250
|
|
|
251 Title : _read_fasta
|
|
|
252 Usage : obj->_read_fasta($file)
|
|
|
253 Function: Internal(not to be used directly)
|
|
|
254 Returns :
|
|
|
255 Args :
|
|
|
256
|
|
|
257
|
|
|
258 =cut
|
|
|
259
|
|
|
260 sub _read_fasta {
|
|
|
261 local (*FILE) = @_;
|
|
|
262 my( $id , $seq , %name2seq);#bala
|
|
|
263 while (<FILE>) {
|
|
|
264 chomp;#bala
|
|
|
265 if (/^>(\S+)/) {
|
|
|
266
|
|
|
267 my $new_id = $1;
|
|
|
268 if ($id) {
|
|
|
269 $name2seq{$id} = $seq;
|
|
|
270 }
|
|
|
271 $id = $new_id ; $seq = "" ;
|
|
|
272 }
|
|
|
273 elsif (eof) {
|
|
|
274 if ($id) {
|
|
|
275 $seq .= $_ ;#bala line instead of $_
|
|
|
276 $name2seq{$id} = $seq;
|
|
|
277 }
|
|
|
278 }
|
|
|
279 else {
|
|
|
280 $seq .= $_
|
|
|
281 }
|
|
|
282 }
|
|
|
283 return %name2seq;
|
|
|
284 }
|
|
|
285
|
|
|
286
|
|
|
287
|
|
|
288 1;
|
|
|
289
|
|
|
290
|
