Bio/Tools/MZEF.pm

   1 #
   2 # BioPerl module for Bio::Tools::MZEF
   3 #
   4 # Please direct questions and support issues to <bioperl-l@bioperl.org>
   5 #
   6 # Cared for by Hilmar Lapp <hlapp-at-gmx.net>
   7 #
   8 # Copyright Hilmar Lapp
   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::Tools::MZEF - Results of one MZEF run
  17
  18 =head1 SYNOPSIS
  19
  20    $mzef = Bio::Tools::MZEF->new(-file => 'result.mzef');
  21    # filehandle:
  22    $mzef = Bio::Tools::MZEF->new( -fh  => \*INPUT );
  23    # to indicate that the sequence was reversed prior to feeding it to MZEF
  24    # and that you want to have this reflected in the strand() attribute of
  25    # the exons, as well have the coordinates translated to the non-reversed
  26    # sequence
  27    $mzef = Bio::Tools::MZEF->new( -file   => 'result.mzef',
  28                                   -strand => -1 );
  29
  30    # parse the results
  31    # note: this class is-a Bio::Tools::AnalysisResult which implements
  32    # Bio::SeqAnalysisParserI, i.e., $genscan->next_feature() is the same
  33    while($gene = $mzef->next_prediction()) {
  34        # $gene is an instance of Bio::Tools::Prediction::Gene
  35
  36        # $gene->exons() returns an array of
  37        # Bio::Tools::Prediction::Exon objects
  38        # all exons:
  39        @exon_arr = $gene->exons();
  40
  41        # internal exons only
  42        @intrl_exons = $gene->exons('Internal');
  43        # note that presently MZEF predicts only internal exons!
  44    }
  45
  46    # essential if you gave a filename at initialization (otherwise the file
  47    # will stay open)
  48    $mzef->close();
  49
  50 =head1 DESCRIPTION
  51
  52 The MZEF module provides a parser for MZEF gene structure prediction
  53 output.
  54
  55 This module inherits off L<Bio::Tools::AnalysisResult> and therefore
  56 implements L<Bio::SeqAnalysisParserI>.
  57
  58 =head1 FEEDBACK
  59
  60 =head2 Mailing Lists
  61
  62 User feedback is an integral part of the evolution of this and other
  63 Bioperl modules. Send your comments and suggestions preferably to one
  64 of the Bioperl mailing lists.  Your participation is much appreciated.
  65
  66   bioperl-l@bioperl.org                  - General discussion
  67   http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
  68
  69 =head2 Support
  70
  71 Please direct usage questions or support issues to the mailing list:
  72
  73 I<bioperl-l@bioperl.org>
  74
  75 rather than to the module maintainer directly. Many experienced and
  76 reponsive experts will be able look at the problem and quickly
  77 address it. Please include a thorough description of the problem
  78 with code and data examples if at all possible.
  79
  80 =head2 Reporting Bugs
  81
  82 Report bugs to the Bioperl bug tracking system to help us keep track
  83 the bugs and their resolution.  Bug reports can be submitted via the
  84 web:
  85
  86   https://github.com/bioperl/bioperl-live/issues
  87
  88 =head1 AUTHOR - Hilmar Lapp
  89
  90 Email hlapp-at-gmx.net (or hilmar.lapp-at-pharma.novartis.com)
  91
  92 =head1 APPENDIX
  93
  94 The rest of the documentation details each of the object
  95 methods. Internal methods are usually preceded with a _
  96
  97 =cut
  98
  99
 100 # Let the code begin...
 101
 102
 103 package Bio::Tools::MZEF;
 104 use strict;
 105
 106 use Bio::Tools::Prediction::Gene;
 107 use Bio::Tools::Prediction::Exon;
 108
 109 use base qw(Bio::Tools::AnalysisResult);
 110
 111 sub _initialize_state {
 112     my($self,@args) = @_;
 113
 114     # first call the inherited method!
 115     my $make = $self->SUPER::_initialize_state(@args);
 116
 117     # handle our own parameters
 118     my ($strand, $params) =
 119         $self->_rearrange([qw(STRAND
 120                               )],
 121                           @args);
 122
 123     # our private state variables
 124     $strand = 1 unless defined($strand);
 125     $self->{'_strand'} = $strand;
 126     $self->{'_preds_parsed'} = 0;
 127     $self->{'_has_cds'} = 0;
 128     # array of pre-parsed predictions
 129     $self->{'_preds'} = [];
 130 }
 131
 132 =head2 analysis_method
 133
 134  Usage     : $mzef->analysis_method();
 135  Purpose   : Inherited method. Overridden to ensure that the name matches
 136              /mzef/i.
 137  Returns   : String
 138  Argument  : n/a
 139
 140 =cut
 141
 142 #-------------
 143 sub analysis_method {
 144 #-------------
 145     my ($self, $method) = @_;
 146     if($method && ($method !~ /mzef/i)) {
 147         $self->throw("method $method not supported in " . ref($self));
 148     }
 149     return $self->SUPER::analysis_method($method);
 150 }
 151
 152 =head2 next_feature
 153
 154  Title   : next_feature
 155  Usage   : while($gene = $mzef->next_feature()) {
 156                   # do something
 157            }
 158  Function: Returns the next gene structure prediction of the MZEF result
 159            file. Call this method repeatedly until FALSE is returned.
 160
 161            The returned object is actually a SeqFeatureI implementing object.
 162            This method is required for classes implementing the
 163            SeqAnalysisParserI interface, and is merely an alias for
 164            next_prediction() at present.
 165
 166            Note that with the present version of MZEF there will only be one
 167            object returned, because MZEF does not predict individual genes
 168            but just potential internal exons.
 169  Example :
 170  Returns : A Bio::Tools::Prediction::Gene object.
 171  Args    :
 172
 173 =cut
 174
 175 sub next_feature {
 176     my ($self,@args) = @_;
 177     # even though next_prediction doesn't expect any args (and this method
 178     # does neither), we pass on args in order to be prepared if this changes
 179     # ever
 180     return $self->next_prediction(@args);
 181 }
 182
 183 =head2 next_prediction
 184
 185  Title   : next_prediction
 186  Usage   : while($gene = $mzef->next_prediction()) {
 187                   # do something
 188            }
 189  Function: Returns the next gene structure prediction of the MZEF result
 190            file. Call this method repeatedly until FALSE is returned.
 191
 192            Note that with the present version of MZEF there will only be one
 193            object returned, because MZEF does not predict individual genes
 194            but just potential internal exons.
 195  Example :
 196  Returns : A Bio::Tools::Prediction::Gene object.
 197  Args    :
 198
 199 =cut
 200
 201 sub next_prediction {
 202     my ($self) = @_;
 203     my $gene;
 204
 205     # if the prediction section hasn't been parsed yet, we do this now
 206     $self->_parse_predictions() unless $self->_predictions_parsed();
 207
 208     # return the next gene structure (transcript)
 209     return $self->_prediction();
 210 }
 211
 212 =head2 _parse_predictions
 213
 214  Title   : _parse_predictions()
 215  Usage   : $obj->_parse_predictions()
 216  Function: Parses the prediction section. Automatically called by
 217            next_prediction() if not yet done.
 218  Example :
 219  Returns :
 220
 221 =cut
 222
 223 sub _parse_predictions {
 224     my ($self) = @_;
 225     my ($method); # set but not used presently
 226     my $exon_tag = "InternalExon";
 227     my $gene;
 228     # my $seqname; # name given in output is poorly formatted
 229     my $seqlen;
 230     my $prednr = 1;
 231
 232     while(defined($_ = $self->_readline())) {
 233         if(/^\s*(\d+)\s*-\s*(\d+)\s+/) {
 234             # exon or signal
 235             if(! defined($gene)) {
 236                 $gene = Bio::Tools::Prediction::Gene->new(
 237                                        '-primary' => "GenePrediction$prednr",
 238                                        '-source' => 'MZEF');
 239             }
 240             # we handle start-end first because may not be space delimited
 241             # for large numbers
 242             my ($start,$end) = ($1,$2);
 243             s/^\s*(\d+)\s*-\s*(\d+)\s+//;
 244             # split the rest into fields
 245             chomp();
 246             # format: Coordinates P Fr1 Fr2 Fr3 Orf 3ss Cds 5ss
 247             # index:              0   1   2   3   4   5   6   7
 248             my @flds = split(' ', $_);
 249             # create the feature object depending on the type of signal --
 250             # which is always an (internal) exon for MZEF
 251             my $predobj = Bio::Tools::Prediction::Exon->new();
 252             # set common fields
 253             $predobj->source_tag('MZEF');
 254             $predobj->significance($flds[0]);
 255             $predobj->score($flds[0]); # what shall we set as overall score?
 256             $predobj->strand($self->{'_strand'}); # MZEF searches only one
 257             if($predobj->strand() == 1) {
 258                 $predobj->start($start);
 259                 $predobj->end($end);
 260             } else {
 261                 $predobj->start($seqlen-$end+1);
 262                 $predobj->end($seqlen-$start+1);
 263             }
 264             # set scores
 265             $predobj->start_signal_score($flds[5]);
 266             $predobj->end_signal_score($flds[7]);
 267             $predobj->coding_signal_score($flds[6]);
 268             # frame -- we simply extract the one with highest score from the
 269             # orf field, and store the individual scores for now
 270             my $frm = index($flds[4], "1");
 271             $predobj->frame(($frm < 0) ? undef : $frm);
 272             $predobj->primary_tag($exon_tag);
 273             $predobj->is_coding(1);
 274             # add to gene structure (should be done only when start and end
 275             # are set, in order to allow for proper expansion of the range)
 276             $gene->add_exon($predobj);
 277             next;
 278         }
 279         if(/^\s*Internal .*(MZEF)/) {
 280             $self->analysis_method($1);
 281             next;
 282         }
 283         if(/^\s*File_Name:\s+(\S+)\s+Sequence_length:\s+(\d+)/) {
 284             # $seqname = $1; # this is too poor currently (file name truncated
 285                              # to 10 chars) in order to be sensible enough
 286             $seqlen = $2;
 287             next;
 288         }
 289     }
 290     # $gene->seq_id($seqname);
 291     $self->_add_prediction($gene) if defined($gene);
 292     $self->_predictions_parsed(1);
 293 }
 294
 295 =head2 _prediction
 296
 297  Title   : _prediction()
 298  Usage   : $gene = $obj->_prediction()
 299  Function: internal
 300  Example :
 301  Returns :
 302
 303 =cut
 304
 305 sub _prediction {
 306     my ($self) = @_;
 307
 308     return unless(exists($self->{'_preds'}) && @{$self->{'_preds'}});
 309     return shift(@{$self->{'_preds'}});
 310 }
 311
 312 =head2 _add_prediction
 313
 314  Title   : _add_prediction()
 315  Usage   : $obj->_add_prediction($gene)
 316  Function: internal
 317  Example :
 318  Returns :
 319
 320 =cut
 321
 322 sub _add_prediction {
 323     my ($self, $gene) = @_;
 324
 325     if(! exists($self->{'_preds'})) {
 326         $self->{'_preds'} = [];
 327     }
 328     push(@{$self->{'_preds'}}, $gene);
 329 }
 330
 331 =head2 _predictions_parsed
 332
 333  Title   : _predictions_parsed
 334  Usage   : $obj->_predictions_parsed
 335  Function: internal
 336  Example :
 337  Returns : TRUE or FALSE
 338
 339 =cut
 340
 341 sub _predictions_parsed {
 342     my ($self, $val) = @_;
 343
 344     $self->{'_preds_parsed'} = $val if $val;
 345     if(! exists($self->{'_preds_parsed'})) {
 346         $self->{'_preds_parsed'} = 0;
 347     }
 348     return $self->{'_preds_parsed'};
 349 }
 350
 351
 352 1;