|
0
|
1 #
|
|
|
2 # $Id: XEMBL.pm,v 1.3 2002/10/22 07:38:29 lapp Exp $
|
|
|
3 #
|
|
|
4 # BioPerl module for Bio::DB::XEMBL
|
|
|
5 #
|
|
|
6 # Cared for by Lincoln Stein
|
|
|
7 #
|
|
|
8 # Copyright Lincoln Stein
|
|
|
9 #
|
|
|
10 # You may distribute this module under the same terms as perl itself
|
|
|
11
|
|
|
12 # POD documentation - main docs before the code
|
|
|
13
|
|
|
14 =head1 NAME
|
|
|
15
|
|
|
16 Bio::DB::XEMBL - Database object interface for XEMBL entry retrieval
|
|
|
17
|
|
|
18 =head1 SYNOPSIS
|
|
|
19
|
|
|
20 use Bio::DB::XEMBL;
|
|
|
21
|
|
|
22 $embl = new Bio::DB::XEMBL;
|
|
|
23
|
|
|
24 # remember that XEMBL_ID does not equal GenBank_ID!
|
|
|
25 $seq = $embl->get_Seq_by_id('BUM'); # EMBL ID
|
|
|
26 print "cloneid is ", $seq->id, "\n";
|
|
|
27
|
|
|
28 # or changeing to accession number and Fasta format ...
|
|
|
29 $seq = $embl->get_Seq_by_acc('J02231'); # XEMBL ACC
|
|
|
30 print "cloneid is ", $seq->id, "\n";
|
|
|
31
|
|
|
32 # especially when using versions, you better be prepared
|
|
|
33 # in not getting what what want
|
|
|
34 eval {
|
|
|
35 $seq = $embl->get_Seq_by_version('J02231.1'); # XEMBL VERSION
|
|
|
36 }
|
|
|
37 print "cloneid is ", $seq->id, "\n" unless $@;
|
|
|
38
|
|
|
39 my $seqio = $embl->get_Stream_by_batch(['U83300','U83301','U83302']);
|
|
|
40 while( my $clone = $seqio->next_seq ) {
|
|
|
41 print "cloneid is ", $clone->id, "\n";
|
|
|
42 }
|
|
|
43
|
|
|
44 =head1 DESCRIPTION
|
|
|
45
|
|
|
46 Allows the dynamic retrieval of Bio::Seq objects from the XEMBL
|
|
|
47 database. See L<Bio::Seq> for details.
|
|
|
48
|
|
|
49 =head1 FEEDBACK
|
|
|
50
|
|
|
51 =head2 Mailing Lists
|
|
|
52
|
|
|
53 User feedback is an integral part of the evolution of this and other
|
|
|
54 Bioperl modules. Send your comments and suggestions preferably to one
|
|
|
55 of the Bioperl mailing lists. Your participation is much appreciated.
|
|
|
56
|
|
|
57 bioperl-l@bioperl.org - General discussion
|
|
|
58 http://bio.perl.org/MailList.html - About the mailing lists
|
|
|
59
|
|
|
60 =head2 Reporting Bugs
|
|
|
61
|
|
|
62 Report bugs to the Bioperl bug tracking system to help us keep track
|
|
|
63 the bugs and their resolution.
|
|
|
64 Bug reports can be submitted via email or the web:
|
|
|
65
|
|
|
66 bioperl-bugs@bio.perl.org
|
|
|
67 http://bugzilla.bioperl.org/
|
|
|
68
|
|
|
69 =head1 AUTHOR - Lincoln Stein
|
|
|
70
|
|
|
71 Email Lincoln Stein E<lt>lstein@cshl.orgE<gt>
|
|
|
72
|
|
|
73 =head1 APPENDIX
|
|
|
74
|
|
|
75 The rest of the documentation details each of the object
|
|
|
76 methods. Internal methods are usually preceded with a _
|
|
|
77
|
|
|
78 =cut
|
|
|
79
|
|
|
80 # Let the code begin...
|
|
|
81
|
|
|
82 package Bio::DB::XEMBL;
|
|
|
83 use strict;
|
|
|
84 use Bio::DB::RandomAccessI;
|
|
|
85 use Bio::DB::XEMBLService 'getNucSeq';
|
|
|
86 # bsml parser appears broken...
|
|
|
87 use Bio::SeqIO::bsml;
|
|
|
88 use File::Temp 'tempfile';
|
|
|
89 use vars qw(@ISA $MODVERSION);
|
|
|
90
|
|
|
91 @ISA = qw(Bio::DB::RandomAccessI);
|
|
|
92 $MODVERSION = '0.1';
|
|
|
93
|
|
|
94 sub new {
|
|
|
95 my ($class, @args ) = @_;
|
|
|
96 my $self = $class->SUPER::new(@args);
|
|
|
97 return $self;
|
|
|
98 }
|
|
|
99
|
|
|
100 =head2 get_Seq_by_id
|
|
|
101
|
|
|
102 Title : get_Seq_by_id
|
|
|
103 Usage : $seq = $db->get_Seq_by_id('ROA1_HUMAN')
|
|
|
104 Function: Gets a Bio::Seq object by its name
|
|
|
105 Returns : a Bio::Seq object
|
|
|
106 Args : the id (as a string) of a sequence
|
|
|
107 Throws : "id does not exist" exception
|
|
|
108
|
|
|
109
|
|
|
110 =cut
|
|
|
111
|
|
|
112 sub get_Seq_by_id {
|
|
|
113 my ($self,@args) = @_;
|
|
|
114 my $seqio = $self->get_Stream_by_batch([@args]);
|
|
|
115 return $seqio->next_seq;
|
|
|
116 }
|
|
|
117
|
|
|
118 =head2 get_Stream_by_batch
|
|
|
119
|
|
|
120 Title : get_Stream_by_batch
|
|
|
121 Usage : $seq = $db->get_Stream_by_batch($ref);
|
|
|
122 Function: Retrieves Seq objects from XEMBL 'en masse', rather than one
|
|
|
123 at a time. Currently this is not particularly efficient, as
|
|
|
124 it loads the entire result into memory and parses it.
|
|
|
125 Example :
|
|
|
126 Returns : a Bio::SeqIO stream object
|
|
|
127 Args : $ref : an array reference containing a list of unique
|
|
|
128 ids/accession numbers.
|
|
|
129
|
|
|
130 =cut
|
|
|
131
|
|
|
132 sub get_Stream_by_batch {
|
|
|
133 my ($self, $ids) = @_;
|
|
|
134 $self->throw("expected an array ref, but got $ids")
|
|
|
135 unless ref($ids) eq 'ARRAY';
|
|
|
136 my @args = @$ids;
|
|
|
137 my $result = getNucSeq(SOAP::Data->name(format=>'bsml'),
|
|
|
138 SOAP::Data->name(ids=>"@args"))
|
|
|
139 or $self->throw('id does not exist');
|
|
|
140 my($fh,$filename) = tempfile(File::Spec->tmpdir . '/bsmlXXXXXX',SUFFIX=>'.bsml');
|
|
|
141 print $fh $result;
|
|
|
142 close $fh;
|
|
|
143 my $seqio = Bio::SeqIO->new(-file=>$filename,-format=>'bsml');
|
|
|
144 unlink $filename;
|
|
|
145 $seqio;
|
|
|
146 }
|
|
|
147
|
|
|
148 *get_Stream_by_id = \&get_Stream_by_batch;
|
|
|
149
|
|
|
150 =head2 get_Seq_by_acc
|
|
|
151
|
|
|
152 Title : get_Seq_by_acc
|
|
|
153 Usage : $seq = $db->get_Seq_by_acc('X77802');
|
|
|
154 Function: Gets a Bio::Seq object by accession number
|
|
|
155 Returns : A Bio::Seq object
|
|
|
156 Args : accession number (as a string)
|
|
|
157 Throws : "acc does not exist" exception
|
|
|
158
|
|
|
159
|
|
|
160 =cut
|
|
|
161
|
|
|
162 sub get_Seq_by_acc{
|
|
|
163 my ($self,@args) = @_;
|
|
|
164 return $self->get_Seq_by_id(@args);
|
|
|
165 }
|
|
|
166
|
|
|
167 =head2 get_Seq_by_version
|
|
|
168
|
|
|
169 Title : get_Seq_by_version
|
|
|
170 Usage : $seq = $db->get_Seq_by_version('X77802.1');
|
|
|
171 Function: Gets a Bio::Seq object by sequence version
|
|
|
172 Returns : A Bio::Seq object
|
|
|
173 Args : accession.version (as a string)
|
|
|
174 Throws : "acc.version does not exist" exception
|
|
|
175
|
|
|
176 =cut
|
|
|
177
|
|
|
178 sub get_Seq_by_version{
|
|
|
179 my ($self,@args) = @_;
|
|
|
180 return $self->get_Seq_by_id(@args);
|
|
|
181 }
|
|
|
182
|
|
|
183
|
|
|
184 1;
|
