line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
#--------------------------------------------------------- |
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
#ISA SiteMatrix, HAS InstanceSite |
4
|
|
|
|
|
|
|
|
5
|
|
|
|
|
|
|
=head1 NAME |
6
|
|
|
|
|
|
|
|
7
|
|
|
|
|
|
|
Bio::Matrix::PSM::Psm - handle combination of site matricies |
8
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
=head1 SYNOPSIS |
10
|
|
|
|
|
|
|
|
11
|
|
|
|
|
|
|
use Bio::Matrix::PSM::IO; |
12
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
#To get a Psm object from a file use the Psm parser: |
14
|
|
|
|
|
|
|
my $psmIO = Bio::Matrix::PSM::IO->new(-format=>'meme', -file=>$file); |
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
# Now go through all entities in the file with next_psm, which |
17
|
|
|
|
|
|
|
# returns a Psm object see Bio::Matrix::PSM::IO for detailed |
18
|
|
|
|
|
|
|
# documentation (matrix predictions or matrix sequence matches or |
19
|
|
|
|
|
|
|
# both): |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
while (my $psm=$psmIO->next_psm) { |
22
|
|
|
|
|
|
|
my %psm_header=$psm->header; |
23
|
|
|
|
|
|
|
my $ic=$psm_header{IC}; |
24
|
|
|
|
|
|
|
my $sites=$psm_header{sites}; |
25
|
|
|
|
|
|
|
my $width=$psm_header{width}; |
26
|
|
|
|
|
|
|
my $score=$psm_header{e_val}; |
27
|
|
|
|
|
|
|
my $IUPAC=$psm->IUPAC; |
28
|
|
|
|
|
|
|
my $instances=$psm->instances; |
29
|
|
|
|
|
|
|
foreach my $instance (@{$instances}) { |
30
|
|
|
|
|
|
|
my $id=$instance->primary_id; |
31
|
|
|
|
|
|
|
#Do something with the id |
32
|
|
|
|
|
|
|
} |
33
|
|
|
|
|
|
|
} |
34
|
|
|
|
|
|
|
|
35
|
|
|
|
|
|
|
#or create from memmory: |
36
|
|
|
|
|
|
|
my $psm= Bio::Matrix::PSM::Psm->new( -pA=>\@pA,-pC=>\@pC,-pG=>\@pG,-pT=>\@pT, |
37
|
|
|
|
|
|
|
-id=>$id, |
38
|
|
|
|
|
|
|
-instances=>$instances, -e_val=>$e_val, |
39
|
|
|
|
|
|
|
-IC=>$ic, -width=>$width, -sites=>$sites) |
40
|
|
|
|
|
|
|
|
41
|
|
|
|
|
|
|
# where pA through pG are the respective frequencies of the matrix (see also |
42
|
|
|
|
|
|
|
# Bio::Matrix::PSM::SiteMatrix), and everything else is self-explenatory, |
43
|
|
|
|
|
|
|
# except for -instances (reference to an array of |
44
|
|
|
|
|
|
|
# Bio::Matrix::PSM::InstanceSite objects) which is documented bellow. |
45
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
=head1 DESCRIPTION |
47
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
To handle a combination of site matrices and/or their corresponding |
49
|
|
|
|
|
|
|
sequence matches (instances). This object inherits from |
50
|
|
|
|
|
|
|
Bio::Matrix::PSM::SiteMatrix, so you can use the respective |
51
|
|
|
|
|
|
|
methods. It may hold also an array of Bio::Matrix::PSM::InstanceSite |
52
|
|
|
|
|
|
|
object, but you will have to retrieve these through |
53
|
|
|
|
|
|
|
Bio::Matrix::PSM::Psm-Einstances method (see below). To some extent |
54
|
|
|
|
|
|
|
this is an expanded SiteMatrix object, holding data from analysis that |
55
|
|
|
|
|
|
|
also deal with sequence matches of a particular matrix. |
56
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
|
58
|
|
|
|
|
|
|
=head2 DESIGN ISSUES |
59
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
This does not make too much sense to me I am mixing PSM with PSM |
61
|
|
|
|
|
|
|
sequence matches Though they are very closely related, I am not |
62
|
|
|
|
|
|
|
satisfied by the way this is implemented here. Heikki suggested |
63
|
|
|
|
|
|
|
different objects when one has something like meme But does this mean |
64
|
|
|
|
|
|
|
we have to write a different objects for mast, meme, transfac, |
65
|
|
|
|
|
|
|
theiresias, etc.? To me the best way is to return SiteMatrix object + |
66
|
|
|
|
|
|
|
arrray of InstanceSite objects and then mast will return undef for |
67
|
|
|
|
|
|
|
SiteMatrix and transfac will return undef for InstanceSite. Probably I |
68
|
|
|
|
|
|
|
cannot see some other design issues that might arise from such |
69
|
|
|
|
|
|
|
approach, but it seems more straightforward. Hilmar does not like |
70
|
|
|
|
|
|
|
this beacause it is an exception from the general BioPerl rules Should |
71
|
|
|
|
|
|
|
I leave this as an option? Also the header rightfully belongs the |
72
|
|
|
|
|
|
|
driver object, and could be retrieved as hashes. I do not think it |
73
|
|
|
|
|
|
|
can be done any other way, unless we want to create even one more |
74
|
|
|
|
|
|
|
object with very unclear content. |
75
|
|
|
|
|
|
|
|
76
|
|
|
|
|
|
|
=head1 FEEDBACK |
77
|
|
|
|
|
|
|
|
78
|
|
|
|
|
|
|
=head2 Mailing Lists |
79
|
|
|
|
|
|
|
|
80
|
|
|
|
|
|
|
User feedback is an integral part of the evolution of this |
81
|
|
|
|
|
|
|
and other Bioperl modules. Send your comments and suggestions preferably |
82
|
|
|
|
|
|
|
to one of the Bioperl mailing lists. |
83
|
|
|
|
|
|
|
Your participation is much appreciated. |
84
|
|
|
|
|
|
|
|
85
|
|
|
|
|
|
|
bioperl-l@bioperl.org - General discussion |
86
|
|
|
|
|
|
|
http://bioperl.org/wiki/Mailing_lists - About the mailing lists |
87
|
|
|
|
|
|
|
|
88
|
|
|
|
|
|
|
=head2 Support |
89
|
|
|
|
|
|
|
|
90
|
|
|
|
|
|
|
Please direct usage questions or support issues to the mailing list: |
91
|
|
|
|
|
|
|
|
92
|
|
|
|
|
|
|
I |
93
|
|
|
|
|
|
|
|
94
|
|
|
|
|
|
|
rather than to the module maintainer directly. Many experienced and |
95
|
|
|
|
|
|
|
reponsive experts will be able look at the problem and quickly |
96
|
|
|
|
|
|
|
address it. Please include a thorough description of the problem |
97
|
|
|
|
|
|
|
with code and data examples if at all possible. |
98
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
=head2 Reporting Bugs |
100
|
|
|
|
|
|
|
|
101
|
|
|
|
|
|
|
Report bugs to the Bioperl bug tracking system to help us keep track |
102
|
|
|
|
|
|
|
the bugs and their resolution. Bug reports can be submitted via the |
103
|
|
|
|
|
|
|
web: |
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
https://github.com/bioperl/bioperl-live/issues |
106
|
|
|
|
|
|
|
|
107
|
|
|
|
|
|
|
=head1 AUTHOR - Stefan Kirov |
108
|
|
|
|
|
|
|
|
109
|
|
|
|
|
|
|
Email skirov@utk.edu |
110
|
|
|
|
|
|
|
|
111
|
|
|
|
|
|
|
|
112
|
|
|
|
|
|
|
=head1 DISCLAIMER |
113
|
|
|
|
|
|
|
|
114
|
|
|
|
|
|
|
This software is provided "as is" without warranty of any kind. |
115
|
|
|
|
|
|
|
|
116
|
|
|
|
|
|
|
=head1 SEE ALSO |
117
|
|
|
|
|
|
|
|
118
|
|
|
|
|
|
|
SiteMatrix, meme, transfac, InstanceSite |
119
|
|
|
|
|
|
|
|
120
|
|
|
|
|
|
|
=head1 APPENDIX |
121
|
|
|
|
|
|
|
|
122
|
|
|
|
|
|
|
=cut |
123
|
|
|
|
|
|
|
|
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
# Let the code begin... |
126
|
|
|
|
|
|
|
package Bio::Matrix::PSM::Psm; |
127
|
3
|
|
|
3
|
|
336
|
use Bio::Matrix::PSM::InstanceSite; |
|
3
|
|
|
|
|
5
|
|
|
3
|
|
|
|
|
79
|
|
128
|
3
|
|
|
3
|
|
10
|
use strict; |
|
3
|
|
|
|
|
5
|
|
|
3
|
|
|
|
|
62
|
|
129
|
|
|
|
|
|
|
|
130
|
3
|
|
|
3
|
|
11
|
use base qw(Bio::Matrix::PSM::SiteMatrix Bio::Matrix::PSM::PsmI Bio::Annotation::Collection); |
|
3
|
|
|
|
|
3
|
|
|
3
|
|
|
|
|
1430
|
|
131
|
|
|
|
|
|
|
|
132
|
|
|
|
|
|
|
@Bio::Matrix::PSM::Psm::HEADER = qw(e_val sites IC width); |
133
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
=head2 new |
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
Title : new |
137
|
|
|
|
|
|
|
Usage : my $psm= Bio::Matrix::PSM::Psm->new( -pA=>\@pA,-pC=>\@pC, |
138
|
|
|
|
|
|
|
-pG=>\@pG,-pT=>\@pT,-id=>$id, |
139
|
|
|
|
|
|
|
-instances=>$instances, |
140
|
|
|
|
|
|
|
-e_val=>$e_val, |
141
|
|
|
|
|
|
|
-IC=>$ic, -width=>$width, |
142
|
|
|
|
|
|
|
-sites=>$sites) |
143
|
|
|
|
|
|
|
Function: Creates a new Bio::Matrix::PSM::Psm object |
144
|
|
|
|
|
|
|
Throws : |
145
|
|
|
|
|
|
|
Example : |
146
|
|
|
|
|
|
|
Returns : Bio::Matrix::PSM::Psm object |
147
|
|
|
|
|
|
|
Args : hash |
148
|
|
|
|
|
|
|
|
149
|
|
|
|
|
|
|
|
150
|
|
|
|
|
|
|
=cut |
151
|
|
|
|
|
|
|
|
152
|
|
|
|
|
|
|
sub new { |
153
|
5
|
|
|
5
|
1
|
22
|
my ($caller,@args) = @_; |
154
|
5
|
|
33
|
|
|
25
|
my $class = ref($caller) || $caller; |
155
|
5
|
|
|
|
|
36
|
my $self = $class->SUPER::new(@args); |
156
|
5
|
|
|
|
|
11
|
$self->{'_annotation'} = {}; #Init from Annotation::Collection |
157
|
5
|
|
|
|
|
46
|
$self->_typemap(Bio::Annotation::TypeManager->new()); #same |
158
|
5
|
|
|
|
|
24
|
($self->{instances})=$self->_rearrange(['INSTANCES'], @args); |
159
|
5
|
|
|
|
|
15
|
return $self; |
160
|
|
|
|
|
|
|
} |
161
|
|
|
|
|
|
|
|
162
|
|
|
|
|
|
|
|
163
|
|
|
|
|
|
|
=head2 instances |
164
|
|
|
|
|
|
|
|
165
|
|
|
|
|
|
|
Title : instances |
166
|
|
|
|
|
|
|
Usage : my @instances=@{$psm->instances}; |
167
|
|
|
|
|
|
|
Function: Gets/sets the instances (Bio::Matrix::PSM::InstanceSite objects) |
168
|
|
|
|
|
|
|
associated with the Psm object |
169
|
|
|
|
|
|
|
Throws : |
170
|
|
|
|
|
|
|
Example : |
171
|
|
|
|
|
|
|
Returns : array reference (Bio::Matrix::PSM::InstanceSite objects) |
172
|
|
|
|
|
|
|
Args : array reference (Bio::Matrix::PSM::InstanceSite objects) |
173
|
|
|
|
|
|
|
|
174
|
|
|
|
|
|
|
=cut |
175
|
|
|
|
|
|
|
|
176
|
|
|
|
|
|
|
sub instances { |
177
|
2
|
|
|
2
|
1
|
5
|
my $self = shift; |
178
|
2
|
|
|
|
|
2
|
my $prev = $self->{instances}; |
179
|
2
|
50
|
|
|
|
8
|
if (@_) { $self->{instances} = shift; } |
|
0
|
|
|
|
|
0
|
|
180
|
2
|
|
|
|
|
4
|
return $prev; |
181
|
|
|
|
|
|
|
} |
182
|
|
|
|
|
|
|
|
183
|
|
|
|
|
|
|
|
184
|
|
|
|
|
|
|
=head2 header |
185
|
|
|
|
|
|
|
|
186
|
|
|
|
|
|
|
Title : header |
187
|
|
|
|
|
|
|
Usage : my %header=$psm->header; |
188
|
|
|
|
|
|
|
my $ic=$psm->header('IC'); |
189
|
|
|
|
|
|
|
Function: Gets the general information, common for most files, |
190
|
|
|
|
|
|
|
dealing with PSM such as information content (IC), score |
191
|
|
|
|
|
|
|
(e-value, etc.), number of sites (sites) and width. This |
192
|
|
|
|
|
|
|
list may expand. The current list should be in |
193
|
|
|
|
|
|
|
@Bio::Matrix::PSM::Psm::HEADER. Returns undef if an |
194
|
|
|
|
|
|
|
argument is supplied that is not in |
195
|
|
|
|
|
|
|
@Bio::Matrix::PSM::meme::HEADER. |
196
|
|
|
|
|
|
|
Throws : |
197
|
|
|
|
|
|
|
Example : |
198
|
|
|
|
|
|
|
Returns : hash or string |
199
|
|
|
|
|
|
|
Args : string (IC, e_val...) |
200
|
|
|
|
|
|
|
|
201
|
|
|
|
|
|
|
=cut |
202
|
|
|
|
|
|
|
|
203
|
|
|
|
|
|
|
sub header { |
204
|
2
|
|
|
2
|
1
|
569
|
my $self = shift; |
205
|
2
|
50
|
|
|
|
6
|
return if ($self->{end}); |
206
|
2
|
|
|
|
|
3
|
my %header; |
207
|
2
|
100
|
|
|
|
5
|
if (@_) {my $key=shift; return $self->{$key}; } |
|
1
|
|
|
|
|
2
|
|
|
1
|
|
|
|
|
4
|
|
208
|
1
|
|
|
|
|
2
|
foreach my $key (@Bio::Matrix::PSM::Psm::HEADER) { |
209
|
4
|
|
|
|
|
7
|
$header{$key}=$self->{$key}; |
210
|
|
|
|
|
|
|
} |
211
|
1
|
|
|
|
|
6
|
return %header; |
212
|
|
|
|
|
|
|
} |
213
|
|
|
|
|
|
|
|
214
|
|
|
|
|
|
|
|
215
|
|
|
|
|
|
|
=head2 matrix |
216
|
|
|
|
|
|
|
|
217
|
|
|
|
|
|
|
Title : matrix |
218
|
|
|
|
|
|
|
Usage : my $matrix=$psm->matrix; |
219
|
|
|
|
|
|
|
Function: Gets/sets the SiteMatrix related information |
220
|
|
|
|
|
|
|
Throws : |
221
|
|
|
|
|
|
|
Example : |
222
|
|
|
|
|
|
|
Returns : Bio::Matrix::PSM::SiteMatrix objects |
223
|
|
|
|
|
|
|
Args : Bio::Matrix::PSM::SiteMatrix objects |
224
|
|
|
|
|
|
|
|
225
|
|
|
|
|
|
|
=cut |
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
|
228
|
|
|
|
|
|
|
sub matrix { |
229
|
2
|
|
|
2
|
1
|
793
|
my $self = shift; |
230
|
|
|
|
|
|
|
my $prev = Bio::Matrix::PSM::SiteMatrix->new(-pA=>$self->{probA}, |
231
|
|
|
|
|
|
|
-pC=>$self->{probC}, |
232
|
|
|
|
|
|
|
-pG=>$self->{probG}, |
233
|
|
|
|
|
|
|
-pT=>$self->{probT}, |
234
|
|
|
|
|
|
|
-lA=>$self->{logA}, |
235
|
|
|
|
|
|
|
-lC=>$self->{logC}, |
236
|
|
|
|
|
|
|
-lG=>$self->{logG}, |
237
|
|
|
|
|
|
|
-lT=>$self->{logT}, |
238
|
|
|
|
|
|
|
-IC=>$self->{IC}, |
239
|
|
|
|
|
|
|
-e_val=>$self->{e_val}, |
240
|
2
|
|
|
|
|
20
|
-id=>$self->{id}); |
241
|
2
|
100
|
|
|
|
7
|
if (@_) { |
242
|
1
|
|
|
|
|
2
|
my $matrix=shift; |
243
|
1
|
|
|
|
|
5
|
$self->{IC} = $matrix->IC; |
244
|
1
|
|
|
|
|
2
|
$self->{probA}=$matrix->{probA}; |
245
|
1
|
|
|
|
|
1
|
$self->{probC}=$matrix->{probC}; |
246
|
1
|
|
|
|
|
2
|
$self->{probG}=$matrix->{probG}; |
247
|
1
|
|
|
|
|
1
|
$self->{probT}=$matrix->{probT}; |
248
|
1
|
|
|
|
|
5
|
$self->{e_val}=$matrix->e_val; |
249
|
1
|
|
|
|
|
4
|
$self->{id}=$matrix->id; |
250
|
|
|
|
|
|
|
} |
251
|
2
|
|
|
|
|
14
|
return $prev; |
252
|
|
|
|
|
|
|
} |
253
|
|
|
|
|
|
|
|
254
|
|
|
|
|
|
|
1; |
255
|
|
|
|
|
|
|
|