line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
# BioPerl module for Bio::Tools::Run::Genscan |
2
|
|
|
|
|
|
|
# |
3
|
|
|
|
|
|
|
# Please direct questions and support issues to |
4
|
|
|
|
|
|
|
# |
5
|
|
|
|
|
|
|
# Cared for by |
6
|
|
|
|
|
|
|
# |
7
|
|
|
|
|
|
|
# Copyright Balamurugan Kumarasamy |
8
|
|
|
|
|
|
|
# |
9
|
|
|
|
|
|
|
# You may distribute this module under the same terms as perl itself |
10
|
|
|
|
|
|
|
# POD documentation - main docs before the code |
11
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
=head1 NAME |
13
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
Bio::Tools::Run::Genscan - Object for identifying genes in a |
15
|
|
|
|
|
|
|
given sequence given a matrix(for appropriate organisms). |
16
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
=head1 SYNOPSIS |
18
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
# Build a Genscan factory |
20
|
|
|
|
|
|
|
my $param = ('MATRIX'=>HumanIso.smat); |
21
|
|
|
|
|
|
|
my $factory = Bio::Tools::Run::Genscan->new($param); |
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
# Pass the factory a Bio::Seq object |
24
|
|
|
|
|
|
|
#@genes is an array of Bio::Tools::Predictions::Gene objects |
25
|
|
|
|
|
|
|
my @genes = $factory->run($seq); |
26
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
=head1 DESCRIPTION |
28
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
Genscan is a gene identifying program developed by Christopher Burge |
30
|
|
|
|
|
|
|
http://genes.mit.edu/burgelab/ |
31
|
|
|
|
|
|
|
|
32
|
|
|
|
|
|
|
By default it looks for an executable called I and data/parameter files |
33
|
|
|
|
|
|
|
in the directory specified by the I environmental variable. |
34
|
|
|
|
|
|
|
|
35
|
|
|
|
|
|
|
=head1 FEEDBACK |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
=head2 Mailing Lists |
38
|
|
|
|
|
|
|
|
39
|
|
|
|
|
|
|
User feedback is an integral part of the evolution of this and other |
40
|
|
|
|
|
|
|
Bioperl modules. Send your comments and suggestions preferably to one |
41
|
|
|
|
|
|
|
of the Bioperl mailing lists. Your participation is much appreciated. |
42
|
|
|
|
|
|
|
|
43
|
|
|
|
|
|
|
bioperl-l@bioperl.org - General discussion |
44
|
|
|
|
|
|
|
http://bioperl.org/wiki/Mailing_lists - About the mailing lists |
45
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
=head2 Support |
47
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
Please direct usage questions or support issues to the mailing list: |
49
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
I |
51
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
rather than to the module maintainer directly. Many experienced and |
53
|
|
|
|
|
|
|
reponsive experts will be able look at the problem and quickly |
54
|
|
|
|
|
|
|
address it. Please include a thorough description of the problem |
55
|
|
|
|
|
|
|
with code and data examples if at all possible. |
56
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
=head2 Reporting Bugs |
58
|
|
|
|
|
|
|
|
59
|
|
|
|
|
|
|
Report bugs to the Bioperl bug tracking system to help us keep track |
60
|
|
|
|
|
|
|
the bugs and their resolution. Bug reports can be submitted via the |
61
|
|
|
|
|
|
|
web: |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
http://redmine.open-bio.org/projects/bioperl/ |
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
=head1 AUTHOR - Bala |
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
Email savikalpa@fugu-sg.org |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
=head1 APPENDIX |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
The rest of the documentation details each of the object |
72
|
|
|
|
|
|
|
methods. Internal methods are usually preceded with a _ |
73
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
=cut |
75
|
|
|
|
|
|
|
|
76
|
|
|
|
|
|
|
package Bio::Tools::Run::Genscan; |
77
|
|
|
|
|
|
|
|
78
|
1
|
|
|
|
|
59
|
use vars qw($AUTOLOAD @ISA $PROGRAM $PROGRAMDIR |
79
|
1
|
|
|
1
|
|
98088
|
$PROGRAMNAME @GENSCAN_PARAMS %OK_FIELD); |
|
1
|
|
|
|
|
2
|
|
80
|
1
|
|
|
1
|
|
4
|
use strict; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
14
|
|
81
|
1
|
|
|
1
|
|
546
|
use Bio::Seq; |
|
1
|
|
|
|
|
42700
|
|
|
1
|
|
|
|
|
27
|
|
82
|
1
|
|
|
1
|
|
467
|
use Bio::SeqIO; |
|
1
|
|
|
|
|
18761
|
|
|
1
|
|
|
|
|
24
|
|
83
|
1
|
|
|
1
|
|
4
|
use Bio::Root::Root; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
13
|
|
84
|
1
|
|
|
1
|
|
2
|
use Bio::Root::IO; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
13
|
|
85
|
1
|
|
|
1
|
|
381
|
use Bio::Factory::ApplicationFactoryI; |
|
1
|
|
|
|
|
113
|
|
|
1
|
|
|
|
|
24
|
|
86
|
1
|
|
|
1
|
|
448
|
use Bio::Tools::Genscan; |
|
1
|
|
|
|
|
24221
|
|
|
1
|
|
|
|
|
31
|
|
87
|
1
|
|
|
1
|
|
442
|
use Bio::Tools::Run::WrapperBase; |
|
1
|
|
|
|
|
3
|
|
|
1
|
|
|
|
|
53
|
|
88
|
|
|
|
|
|
|
|
89
|
|
|
|
|
|
|
@ISA = qw(Bio::Root::Root Bio::Tools::Run::WrapperBase); |
90
|
|
|
|
|
|
|
|
91
|
|
|
|
|
|
|
BEGIN { |
92
|
1
|
|
|
1
|
|
2
|
@GENSCAN_PARAMS=qw(MATRIX VERBOSE QUIET); |
93
|
1
|
|
|
|
|
3
|
foreach my $attr ( @GENSCAN_PARAMS) |
94
|
3
|
|
|
|
|
444
|
{ $OK_FIELD{$attr}++; } |
95
|
|
|
|
|
|
|
} |
96
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
=head2 program_name |
98
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
Title : program_name |
100
|
|
|
|
|
|
|
Usage : $factory>program_name() |
101
|
|
|
|
|
|
|
Function: holds the program name |
102
|
|
|
|
|
|
|
Returns: string |
103
|
|
|
|
|
|
|
Args : None |
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
=cut |
106
|
|
|
|
|
|
|
|
107
|
|
|
|
|
|
|
sub program_name { |
108
|
0
|
|
|
0
|
1
|
|
return 'genscan'; |
109
|
|
|
|
|
|
|
} |
110
|
|
|
|
|
|
|
|
111
|
|
|
|
|
|
|
=head2 program_dir |
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
Title : program_dir |
114
|
|
|
|
|
|
|
Usage : $factory->program_dir(@params) |
115
|
|
|
|
|
|
|
Function: returns the program directory, obtained from ENV variable. |
116
|
|
|
|
|
|
|
Returns: string |
117
|
|
|
|
|
|
|
Args : |
118
|
|
|
|
|
|
|
|
119
|
|
|
|
|
|
|
=cut |
120
|
|
|
|
|
|
|
|
121
|
|
|
|
|
|
|
sub program_dir { |
122
|
0
|
|
|
0
|
1
|
|
return Bio::Root::IO->catfile($ENV{GENSCANDIR}); |
123
|
|
|
|
|
|
|
} |
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
|
126
|
|
|
|
|
|
|
sub AUTOLOAD { |
127
|
0
|
|
|
0
|
|
|
my $self = shift; |
128
|
0
|
|
|
|
|
|
my $attr = $AUTOLOAD; |
129
|
0
|
|
|
|
|
|
$attr =~ s/.*:://; |
130
|
0
|
|
|
|
|
|
$attr = uc $attr; |
131
|
0
|
0
|
|
|
|
|
$self->throw("Unallowed parameter: $attr !") unless $OK_FIELD{$attr}; |
132
|
0
|
0
|
|
|
|
|
$self->{$attr} = shift if @_; |
133
|
0
|
|
|
|
|
|
return $self->{$attr}; |
134
|
|
|
|
|
|
|
} |
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
sub new { |
137
|
0
|
|
|
0
|
1
|
|
my ($class,@args) = @_; |
138
|
0
|
|
|
|
|
|
my $self = $class->SUPER::new(@args); |
139
|
0
|
|
|
|
|
|
my ($attr, $value); |
140
|
0
|
|
|
|
|
|
while (@args) { |
141
|
0
|
|
|
|
|
|
$attr = shift @args; |
142
|
0
|
|
|
|
|
|
$value = shift @args; |
143
|
0
|
0
|
|
|
|
|
next if( $attr =~ /^-/ ); # don't want named parameters |
144
|
0
|
|
|
|
|
|
$self->$attr($value); |
145
|
|
|
|
|
|
|
} |
146
|
0
|
|
|
|
|
|
return $self; |
147
|
|
|
|
|
|
|
} |
148
|
|
|
|
|
|
|
|
149
|
|
|
|
|
|
|
=head2 predict_genes() |
150
|
|
|
|
|
|
|
|
151
|
|
|
|
|
|
|
Title : predict_genes() |
152
|
|
|
|
|
|
|
Usage : DEPRECATED: use $obj->run($seq) instead |
153
|
|
|
|
|
|
|
Function: Runs genscan and creates an array of Genes |
154
|
|
|
|
|
|
|
Returns : An array of Bio::Tools::Prediction::Gene objects |
155
|
|
|
|
|
|
|
Args : A Bio::PrimarySeqI |
156
|
|
|
|
|
|
|
|
157
|
|
|
|
|
|
|
=cut |
158
|
|
|
|
|
|
|
|
159
|
|
|
|
|
|
|
sub predict_genes{ |
160
|
0
|
|
|
0
|
1
|
|
return shift->run(@_); |
161
|
|
|
|
|
|
|
} |
162
|
|
|
|
|
|
|
|
163
|
|
|
|
|
|
|
=head2 run |
164
|
|
|
|
|
|
|
|
165
|
|
|
|
|
|
|
Title : run |
166
|
|
|
|
|
|
|
Usage : $obj->run($seq) |
167
|
|
|
|
|
|
|
Function: Runs genscan and creates an array of Genes |
168
|
|
|
|
|
|
|
Returns : An array of Bio::Tools::Prediction::Gene objects |
169
|
|
|
|
|
|
|
Args : A Bio::PrimarySeqI |
170
|
|
|
|
|
|
|
|
171
|
|
|
|
|
|
|
=cut |
172
|
|
|
|
|
|
|
|
173
|
|
|
|
|
|
|
sub run { |
174
|
0
|
|
|
0
|
1
|
|
my ($self,$seq) = @_; |
175
|
0
|
|
|
|
|
|
my $infile1 = $self->_writeSeqFile($seq); |
176
|
0
|
|
|
|
|
|
$self->_set_input($infile1); |
177
|
0
|
|
|
|
|
|
my @feat = $self->_run(); |
178
|
0
|
|
|
|
|
|
return @feat; |
179
|
|
|
|
|
|
|
} |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
=head2 _run |
182
|
|
|
|
|
|
|
|
183
|
|
|
|
|
|
|
Title : _run |
184
|
|
|
|
|
|
|
Usage : $obj->_run() |
185
|
|
|
|
|
|
|
Function: Internal(not to be used directly) |
186
|
|
|
|
|
|
|
Returns : An array of Bio::Tools::Prediction::Gene objects |
187
|
|
|
|
|
|
|
Args : |
188
|
|
|
|
|
|
|
|
189
|
|
|
|
|
|
|
=cut |
190
|
|
|
|
|
|
|
|
191
|
|
|
|
|
|
|
sub _run { |
192
|
|
|
|
|
|
|
|
193
|
0
|
|
|
0
|
|
|
my ($self) = @_; |
194
|
0
|
|
|
|
|
|
my @genes; |
195
|
|
|
|
|
|
|
my $gene; |
196
|
|
|
|
|
|
|
|
197
|
0
|
|
|
|
|
|
my $str = $self->executable.' '.$self->MATRIX.' '.$self->{'input'}; |
198
|
0
|
0
|
|
|
|
|
if($self->verbose){ |
199
|
0
|
|
|
|
|
|
$str.=" -v "; |
200
|
|
|
|
|
|
|
} |
201
|
0
|
0
|
|
|
|
|
if($self->quiet){ |
202
|
0
|
0
|
|
|
|
|
my $null = ($^O =~ m/mswin/i) ? 'NUL' : '/dev/null'; |
203
|
0
|
|
|
|
|
|
open(STDERR,">$null"); |
204
|
|
|
|
|
|
|
} |
205
|
0
|
0
|
|
|
|
|
unless (open(GENSCAN, "$str |")){ |
206
|
0
|
|
|
|
|
|
$self->warn("Cannot run $str"); |
207
|
|
|
|
|
|
|
} |
208
|
0
|
|
|
|
|
|
close(STDERR); |
209
|
0
|
|
|
|
|
|
my $genScanParser = Bio::Tools::Genscan->new(-fh=> \*GENSCAN); |
210
|
|
|
|
|
|
|
|
211
|
|
|
|
|
|
|
|
212
|
0
|
|
|
|
|
|
while( $gene = $genScanParser->next_prediction()){ |
213
|
0
|
|
|
|
|
|
push(@genes, $gene); |
214
|
|
|
|
|
|
|
} |
215
|
0
|
|
|
|
|
|
$self->cleanup(); |
216
|
0
|
|
|
|
|
|
return @genes; |
217
|
|
|
|
|
|
|
} |
218
|
|
|
|
|
|
|
|
219
|
|
|
|
|
|
|
=head2 _set_input() |
220
|
|
|
|
|
|
|
|
221
|
|
|
|
|
|
|
Title : _set_input |
222
|
|
|
|
|
|
|
Usage : obj->_set_input($matrixFile,$seqFile) |
223
|
|
|
|
|
|
|
Function: Internal(not to be used directly) |
224
|
|
|
|
|
|
|
Returns : |
225
|
|
|
|
|
|
|
Args : |
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
=cut |
228
|
|
|
|
|
|
|
|
229
|
|
|
|
|
|
|
sub _set_input() { |
230
|
0
|
|
|
0
|
|
|
my ($self,$infile1) = @_; |
231
|
0
|
|
|
|
|
|
$self->{'input'}=$infile1; |
232
|
|
|
|
|
|
|
} |
233
|
|
|
|
|
|
|
|
234
|
|
|
|
|
|
|
=head2 _writeSeqFile() |
235
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
Title : _writeSeqFile |
237
|
|
|
|
|
|
|
Usage : obj->_writeSeqFile($seq) |
238
|
|
|
|
|
|
|
Function: Internal(not to be used directly) |
239
|
|
|
|
|
|
|
Returns : |
240
|
|
|
|
|
|
|
Args : |
241
|
|
|
|
|
|
|
|
242
|
|
|
|
|
|
|
=cut |
243
|
|
|
|
|
|
|
|
244
|
|
|
|
|
|
|
|
245
|
|
|
|
|
|
|
sub _writeSeqFile(){ |
246
|
0
|
|
|
0
|
|
|
my ($self,$seq) = @_; |
247
|
0
|
|
|
|
|
|
my ($tfh,$inputfile) = $self->io->tempfile(-dir=>$self->tempdir); |
248
|
0
|
|
|
|
|
|
my $in = Bio::SeqIO->new(-fh => $tfh , '-format' => 'fasta'); |
249
|
0
|
|
|
|
|
|
$in->write_seq($seq); |
250
|
0
|
|
|
|
|
|
$in->close(); |
251
|
0
|
|
|
|
|
|
close($tfh); |
252
|
0
|
|
|
|
|
|
undef $tfh; |
253
|
0
|
|
|
|
|
|
return $inputfile; |
254
|
|
|
|
|
|
|
} |
255
|
|
|
|
|
|
|
|
256
|
|
|
|
|
|
|
1; |