2 # BioPerl module for Bio::SeqIO::gbdriver
4 # Please direct questions and support issues to <bioperl-l@bioperl.org>
6 # Cared for by Bioperl project bioperl-l(at)bioperl.org
8 # Copyright Chris Fields and contributors see AUTHORS section
10 # You may distribute this module under the same terms as perl itself
12 # POD documentation - main docs before the code
16 Bio::SeqIO::gbdriver - GenBank handler-based push parser
20 #It is probably best not to use this object directly, but
21 #rather go through the SeqIO handler:
23 $stream = Bio::SeqIO->new(-file => $filename,
24 -format => 'gbdriver');
26 while ( my $seq = $stream->next_seq() ) {
27 # do something with $seq
32 This object can transform Bio::Seq objects to and from GenBank flat file
33 databases. The key difference between this parser and the tried-and-true
34 Bio::SeqIO::genbank parser is this version separates the parsing and data
35 manipulation into a 'driver' method (next_seq) and separate object handlers
36 which deal with the data passed to it.
40 The main purpose of the driver routine, in this case next_seq(), is to carve out
41 the data into meaningful chunks which are passed along to relevant handlers (see
44 Each chunk of data in the has a NAME tag attached to it, similar to that for XML
45 parsing. This designates the type of data passed (annotation type or seqfeature)
46 and the handler to be called for processing the data.
48 For GenBank annotations, the data is divided up and passed along to handlers
49 according to whether the data is tagged with a field name (i.e. LOCUS) and
50 whether the field name represents 'primary' annotation (in this case, is present
51 at the beginning of the line, such as REFERENCE). If the field is primary, it is
52 assigned to the NAME tag. Field names which aren't primary (have at least 2
53 spaces before the name, like ORGANISM) are appended to the preceding primary
54 field name as additional tags.
56 For feature table data each new feature name signals the beginning of a new
57 chunk of data. 'FEATURES' is attached to NAME, the feature key ('CDS', 'gene',
58 etc) is attached as the PRIMARY_ID, and the location is assigned to it's own tag
59 name (LOCATION). Feature qualifiers are added as additional keys, with multiple
60 keys included in an array.
62 Once a particular event occurs (new primary tag, sequence, end of record), the
63 data is passed along to be processed by a handler or (if no handler is defined)
66 Internally, the hash ref for a representative annotation (here a REFERENCE)
70 'JOURNAL' => 'Unpublished (2003)',
71 'TITLE' => 'The DNA sequence of Homo sapiens',
72 'NAME' => 'REFERENCE',
73 'REFERENCE' => '1 (bases 1 to 10001)',
74 'AUTHORS' => 'International Human Genome Sequencing Consortium.'
77 and a SeqFeature as this:
84 'LOCATION' => 'complement(3024..6641)',
86 'FEATURE_KEY' => 'gene',
87 'gene' => 'LOC127086',
88 'note' => 'Derived by automated computational analysis using
89 gene prediction method: GNOMON.'
92 Note that any driver implementation would suffice as long as it fulfilled the
99 User feedback is an integral part of the evolution of this and other
100 Bioperl modules. Send your comments and suggestions preferably to one
101 of the Bioperl mailing lists. Your participation is much appreciated.
103 bioperl-l@bioperl.org - General discussion
104 http://bioperl.org/wiki/Mailing_lists - About the mailing lists
108 Please direct usage questions or support issues to the mailing list:
110 I<bioperl-l@bioperl.org>
112 rather than to the module maintainer directly. Many experienced and
113 reponsive experts will be able look at the problem and quickly
114 address it. Please include a thorough description of the problem
115 with code and data examples if at all possible.
117 =head2 Reporting Bugs
119 Report bugs to the Bioperl bug tracking system to help us keep track
120 the bugs and their resolution. Bug reports can be submitted via the web:
122 https://github.com/bioperl/bioperl-live/issues
124 =head1 AUTHOR - Bioperl Project
126 bioperl-l at bioperl.org
128 Original author Elia Stupka, elia -at- tigem.it
132 Ewan Birney birney at ebi.ac.uk
133 Jason Stajich jason at bioperl.org
134 Chris Mungall cjm at fruitfly.bdgp.berkeley.edu
135 Lincoln Stein lstein at cshl.org
136 Heikki Lehvaslaiho, heikki at ebi.ac.uk
137 Hilmar Lapp, hlapp at gmx.net
138 Donald G. Jackson, donald.jackson at bms.com
139 James Wasmuth, james.wasmuth at ed.ac.uk
140 Brian Osborne, bosborne at alum.mit.edu
144 The rest of the documentation details each of the object
145 methods. Internal methods are usually preceded with a _
149 # POD is at the end of the module
151 # Let the code begin...
153 package Bio
::SeqIO
::gbdriver
;
157 use Bio
::SeqIO
::Handler
::GenericRichSeqHandler
;
158 use Bio
::Seq
::SeqFactory
;
160 use base
qw(Bio::SeqIO);
162 # map all annotation keys to consistent INSDC-based tags for all handlers
164 my %FTQUAL_NO_QUOTE = map {$_ => 1} qw(
167 cons_splice direction
171 transl_except transl_table
176 # 1) change this to indicate what should be secondary, not primary, which allows
177 # unknown or new stuff to be passed to handler automatically; current behavior
178 # appends unknowns to previous data, which isn't good since it's subtly passing
180 # 2) add mapping details about how to separate data using specific delimiters
183 # Features are the only ones postprocessed for now
184 # Uncomment relevant code in next_seq and add keys as needed...
185 my %POSTPROCESS_DATA = map {$_ => 1} qw
(FEATURES
);
188 my($self,@args) = @_;
190 $self->SUPER::_initialize
(@args);
191 my $handler = $self->_rearrange([qw(HANDLER)],@args);
192 # hash for functions for decoding keys.
193 $handler ?
$self->seqhandler($handler) :
194 $self->seqhandler(Bio
::SeqIO
::Handler
::GenericRichSeqHandler
->new(
195 -format
=> 'genbank',
196 -verbose
=> $self->verbose,
197 -builder
=> $self->sequence_builder
199 if( ! defined $self->sequence_factory ) {
200 $self->sequence_factory(Bio
::Seq
::SeqFactory
->new
201 (-verbose
=> $self->verbose(),
202 -type
=> 'Bio::Seq::RichSeq'));
209 Usage : $seq = $stream->next_seq()
210 Function: returns the next sequence in the stream
211 Returns : Bio::Seq object
216 # at this point there is minimal sequence validation,
217 # but the parser seems to hold up nicely so far...
222 my ($ann, $data, $annkey);
223 my $endrec = my $seenfeat = 0;
226 my $hobj = $self->seqhandler;
227 my $handlers = $self->seqhandler->handler_methods;
228 #$self->debug(Dumper($handlers));
230 while (defined(my $line = $self->_readline)) {
231 next if $line =~ m{^\s*$};
233 # have to catch this at the top of the loop, then exit SEQ loop on //
234 # The reason? The regex match for ann/feat keys also matches some lines
235 # in the sequence; no easy way around it since some feature keys may
236 # start with a number as well
237 if ($ann && $ann eq 'ORIGIN') {
239 while (defined($line)) {
240 last SEQ
if index($line,'//') == 0;
241 $seqdata->{DATA
} .= uc $line;
242 $line = $self->_readline;
244 $seqdata->{DATA
} =~ tr{0-9 \n}{}d;
246 $endrec = 1 if (index($line,'//')==0);
248 if ($line =~ m{^(\s{0,5})(\w
+)\s
+(.*)$}ox
|| $endrec) {
249 ($ann, $data) = ($2, $3);
250 unless ($seenlocus) {
251 $self->throw("No LOCUS found. Not GenBank in my book!")
252 if ($ann ne 'LOCUS');
255 # use the spacer to determine the annotation type
256 my $len = length($1 || '');
258 $annkey = ($len == 0 || $len > 4) ?
'DATA' : $ann;
260 # Push off the previously cached data to the handler
261 # whenever a new primary annotation or seqfeature is found
262 # Note use of $endrec for catching end of record
263 if (($annkey eq 'DATA') && $seqdata) {
264 chomp $seqdata->{DATA
};
265 # postprocessing for some data
266 if ($seqdata->{NAME
} eq 'FEATURES') {
267 $self->_process_features($seqdata)
270 # using handlers directly, slightly faster
271 #my $method = (exists $handlers->{ $seqdata->{NAME} }) ?
272 # ($handlers->{$seqdata->{NAME}}) :
273 # (exists $handlers->{'_DEFAULT_'}) ?
274 # ($handlers->{'_DEFAULT_'}) :
276 #($method) ? ($hobj->$method($seqdata) ) :
277 # $self->debug("No handler defined for ",$seqdata->{NAME},"\n");
279 # using handler methods in the Handler object, more centralized
280 #$self->debug(Dumper($seqdata));
281 $hobj->data_handler($seqdata);
284 last PARSER
if $endrec;
285 # reset for next round
289 $seqdata->{NAME
} = ($len == 0) ?
$ann : # primary ann
290 ($len > 4 ) ?
'FEATURES': # sf feature key
291 $seqdata->{NAME
}; # all rest are sec. ann
292 if ($seqdata->{NAME
} eq 'FEATURES') {
293 $seqdata->{FEATURE_KEY
} = $ann;
295 # throw back to top if seq is found to avoid regex
296 next PARSER
if $ann eq 'ORIGIN';
299 ($data = $line) =~ s{^\s+}{};
302 my $delim = ($seqdata && $seqdata->{NAME
} eq 'FEATURES') ?
"\n" : ' ';
303 $seqdata->{$annkey} .= ($seqdata->{$annkey}) ?
$delim.$data : $data;
305 return $hobj->build_sequence;
311 my ($ann, $data, $annkey);
312 my $endrec = my $seenfeat = 0;
315 my $hobj = $self->seqhandler;
317 while (defined(my $line = $self->_readline)) {
318 next if $line =~ m{^\s*$};
319 # have to catch this at the top of the loop, then exit SEQ loop on //
320 # The reason? The regex match for ann/feat keys also matches some lines
321 # in the sequence; no easy way around it since some feature keys may
322 # start with a number as well
323 if ($ann && $ann eq 'ORIGIN') {
325 while (defined($line)) {
326 last SEQ
if index($line,'//') == 0;
327 $seqdata->{DATA
} .= uc $line;
328 $line = $self->_readline;
330 $seqdata->{DATA
} =~ tr{0-9 \n}{}d;
332 $endrec = 1 if (index($line,'//')==0);
334 if ($line =~ m{^(\s{0,5})(\w
+)\s
+(.*)$}ox
|| $endrec) {
335 ($ann, $data) = ($2, $3);
336 unless ($seenlocus) {
337 $self->throw("No LOCUS found. Not GenBank in my book!")
338 if ($ann ne 'LOCUS');
341 # use the spacer to determine the annotation type
342 my $len = length($1 || '');
344 $annkey = ($len == 0 || $len > 4) ?
'DATA' : $ann;
346 # Push off the previously cached data to the handler
347 # whenever a new primary annotation or seqfeature is found
348 # Note use of $endrec for catching end of record
349 if (($annkey eq 'DATA') && $seqdata) {
350 chomp $seqdata->{DATA
};
351 # postprocessing for some data
352 if ($seqdata->{NAME
} eq 'FEATURES') {
353 $self->_process_features($seqdata)
355 # using handler methods in the Handler object, more centralized
356 $hobj->data_handler($seqdata);
358 last PARSER
if $endrec;
359 # reset for next round
363 $seqdata->{NAME
} = ($len == 0) ?
$ann : # primary ann
364 ($len > 4 ) ?
'FEATURES': # sf feature key
365 $seqdata->{NAME
}; # all rest are sec. ann
366 if ($seqdata->{NAME
} eq 'FEATURES') {
367 $seqdata->{FEATURE_KEY
} = $ann;
369 # throw back to top if seq is found to avoid regex
370 next PARSER
if $ann eq 'ORIGIN';
372 ($data = $line) =~ s{^\s+}{};
375 my $delim = ($seqdata && $seqdata->{NAME
} eq 'FEATURES') ?
"\n" : ' ';
376 $seqdata->{$annkey} .= ($seqdata->{$annkey}) ?
$delim.$data : $data;
383 Usage : $stream->write_seq($seq)
384 Function: writes the $seq object (must be seq) to the stream
385 Returns : 1 for success and 0 for error
386 Args : array of 1 to n Bio::SeqI objects
391 shift->throw("Use Bio::SeqIO::genbank for output");
392 # maybe make a Writer class as well????
398 Usage : $stream->seqhandler($handler)
399 Function: Get/Set the Bio::Seq::HandlerBaseI object
400 Returns : Bio::Seq::HandlerBaseI
401 Args : Bio::Seq::HandlerBaseI
406 my ($self, $handler) = @_;
408 $self->throw("Not a Bio::HandlerBaseI") unless
409 ref($handler) && $handler->isa("Bio::HandlerBaseI");
410 $self->{'_seqhandler'} = $handler;
412 return $self->{'_seqhandler'};
415 #=head2 _process_features
417 # Title : _process_features
418 # Usage : $self->_process_features($seqdata)
419 # Function: Process feature data chunk into usable bits
425 sub _process_features
{
426 my ($self, $seqdata) = @_;
427 my @ftlines = split m{\n}, $seqdata->{DATA
};
428 delete $seqdata->{DATA
};
429 # don't deal with balancing quotes for now; just get rid of them...
430 # Should we worry about checking whether these are balanced
431 # for round-tripping tests?
432 map { s{"}{}g } @ftlines;
433 # all sfs start with the location...
434 my $qual = 'LOCATION';
436 for my $qualdata (@ftlines) {
437 if ($qualdata =~ m{^/([^=]+)=?(.+)?}) {
438 ($qual, $qualdata) = ($1, $2);
439 $qualdata ||= ''; # for those qualifiers that have no data, like 'pseudo'
440 $ct = (exists $seqdata->{$qual}) ?
441 ((ref($seqdata->{$qual})) ?
scalar(@
{ $seqdata->{$qual} }) : 1)
444 my $delim = ($qual eq 'translation' || exists $FTQUAL_NO_QUOTE{$qual}) ?
446 # if more than one, turn into an array ref and append
448 (exists $seqdata->{$qual}) ?
($seqdata->{$qual}.= $delim.$qualdata || '') :
449 ($seqdata->{$qual} .= $qualdata || '');
451 if (!ref($seqdata->{$qual})) {
452 $seqdata->{$qual} = [$seqdata->{$qual}];
454 (exists $seqdata->{$qual}->[$ct]) ?
(($seqdata->{$qual}->[$ct]) .= $delim.$qualdata) :
455 (($seqdata->{$qual}->[$ct]) .= $qualdata);