lib/Bio/Tools/Signalp.pm

   1 # Parser module for Signalp Bio::Tools::Signalp
   2 #
   3 #
   4 # Based on the EnsEMBL module
   5 # Bio::EnsEMBL::Pipeline::Runnable::Protein::Signalp originally
   6 # written by Marc Sohrmann (ms2@sanger.ac.uk) Written in BioPipe by
   7 # Please direct questions and support issues to <bioperl-l@bioperl.org>
   8 #
   9 # Balamurugan Kumarasamy <savikalpa@fugu-sg.org> Cared for by the Fugu
  10 # Informatics team (fuguteam@fugu-sg.org)
  11
  12 # You may distribute this module under the same terms as perl itself
  13 #
  14 # POD documentation - main docs before the code
  15
  16 =head1 NAME
  17
  18 Bio::Tools::Signalp - parser for Signalp output
  19
  20 =head1 SYNOPSIS
  21
  22  use Bio::Tools::Signalp;
  23
  24  my $parser = Bio::Tools::Signalp->new(-fh =>$filehandle );
  25
  26  while( my $sp_feat = $parser->next_result ) {
  27    if ($sp_feat->score > 0.9) {
  28       push @likely_sigpep, $sp_feat;
  29    }
  30  }
  31
  32 =head1 DESCRIPTION
  33
  34 C<SignalP> predicts the presence and location of signal peptide
  35 cleavage sites in amino acid sequences.
  36
  37 L<Bio::Tools::Signalp> parses the output of C<SignalP> to provide a
  38 L<Bio::SeqFeature::Generic> object describing the signal peptide
  39 found, if any. It returns a variety of tags extracted from the NN and HMM
  40 analysis. Most importantly, the C<score()> attribute contains the
  41 NN probability of this being a true signal peptide.
  42
  43
  44 =head1 FEEDBACK
  45
  46 =head2 Mailing Lists
  47
  48  User feedback is an integral part of the evolution of this and other
  49  Bioperl modules. Send your comments and suggestions preferably to
  50  the Bioperl mailing list.  Your participation is much appreciated.
  51
  52   bioperl-l@bioperl.org                  - General discussion
  53   http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
  54
  55 =head2 Support
  56
  57 Please direct usage questions or support issues to the mailing list:
  58
  59 I<bioperl-l@bioperl.org>
  60
  61 rather than to the module maintainer directly. Many experienced and
  62 reponsive experts will be able look at the problem and quickly
  63 address it. Please include a thorough description of the problem
  64 with code and data examples if at all possible.
  65
  66 =head2 Reporting Bugs
  67
  68 Report bugs to the Bioperl bug tracking system to help us keep track
  69 of the bugs and their resolution. Bug reports can be submitted va the
  70 web:
  71
  72   https://github.com/bioperl/bioperl-live/issues
  73
  74 =head1 AUTHOR
  75
  76 # Please direct questions and support issues to I<bioperl-l@bioperl.org>
  77
  78 Based on the EnsEMBL module Bio::EnsEMBL::Pipeline::Runnable::Protein::Signalp
  79 originally written by Marc Sohrmann (ms2_AT_sanger.ac.uk). Written in BioPipe by
  80 Balamurugan Kumarasamy savikalpa_AT_fugu-sg.org. Cared for by the Fugu
  81 Informatics team (fuguteam_AT_fugu-sg.org)
  82
  83 =head1 CONTRIBUTORS
  84
  85 Torsten Seemann - torsten.seemann AT infotech.monash.edu.au
  86
  87 =head1 APPENDIX
  88
  89 The rest of the documentation details each of the object methods.
  90 Internal methods are usually preceded with a _
  91
  92 =cut
  93
  94 package Bio::Tools::Signalp;
  95
  96 use strict;
  97
  98 use Bio::SeqFeature::Generic;
  99 use base qw(Bio::Root::Root Bio::Root::IO);
 100
 101
 102
 103 =head2 new
 104
 105  Title   : new
 106  Usage   : my $obj = Bio::Tools::Signalp->new();
 107  Function: Builds a new Bio::Tools::Signalp object
 108  Returns : Bio::Tools::Signalp
 109  Args    : -fh/-file => $val, # for initing input, see Bio::Root::IO
 110
 111 =cut
 112
 113 sub new {
 114       my($class,@args) = @_;
 115
 116       my $self = $class->SUPER::new(@args);
 117       $self->_initialize_io(@args);
 118
 119       return $self;
 120 }
 121
 122 =head2 next_result
 123
 124  Title   : next_result
 125  Usage   : my $feat = $signalp->next_result
 126  Function: Get the next result set from parser data
 127  Returns : Bio::SeqFeature::Generic
 128  Args    : none
 129
 130 =cut
 131
 132 sub next_result {
 133         my ($self) = @_;
 134
 135         while (my $line=$self->_readline()) {
 136            chomp $line;
 137
 138            if ($line=~/^\>(\S+)/) {
 139               $self->_seqname($1);
 140            }
 141            elsif ($line=~/max\.\s+Y\s+(\S+)\s+\S+\s+\S+\s+(\S+)/) {
 142               $self->_fact1($2);
 143            }
 144            elsif ($line=~/mean\s+S\s+(\S+)\s+\S+\s+\S+\s+(\S+)/) {
 145               my $fact2 = $2;
 146
 147               if ($fact2 eq 'YES' and $self->_fact1 eq 'YES') {
 148
 149                   my $line = $self->_readline();
 150
 151                   ###########################################
 152                   # modification to suit new SignalP output
 153                   ###########################################
 154                                         chomp $line;
 155                                         #print STDERR "********** <$line>\n";
 156                                         if ($line =~ /\s+D\s+.*/) {
 157                                                 $line = $self->_readline();
 158                                         }
 159                                         #print STDERR "********** <$line>\n";
 160                                         my $end;
 161                                   ###########################################
 162
 163
 164                   if ($line =~ /Most likely cleavage site between pos\.\s+(\d+)/) {
 165                       my $end = $1;
 166                       my (%feature);
 167                       $feature{seq_id} = $self->_seqname;
 168                       $feature{start} = 1;
 169                       $feature{end} = $end;
 170                       $feature{source_tag} = 'Signalp';
 171                       $feature{primary}= 'signal_peptide';
 172                       $self->_parse_hmm_result(\%feature);
 173                       my $new_feat = $self->_create_feature (\%feature);
 174                       return $new_feat;
 175                   }
 176                   else {
 177                       $self->throw ("parsing problem in signalp");
 178                   }
 179
 180               }
 181            }
 182
 183         }
 184 }
 185
 186 =head2 _parse_hmm_result
 187
 188  Title   : _parse_hmm_result
 189  Usage   : $self->_parse_hmm_result(\%feature)
 190  Function: Internal (not to be used directly)
 191  Returns : hash of feature values
 192  Args    : hash of more feature values
 193
 194 =cut
 195
 196 sub _parse_hmm_result {
 197     my ($self, $feature_hash) = @_;
 198     while(my $line = $self->_readline){
 199         chomp $line;
 200         if($line =~ /Prediction: (.+)$/){
 201             $feature_hash->{hmmProdiction} = $1;
 202         }elsif($line =~ /Signal peptide probability: ([0-9\.]+)/){
 203             $feature_hash->{peptideProb} = $1;
 204         }elsif($line =~ /Signal anchor probability: ([0-9\.]+)/){
 205             $feature_hash->{anchorProb} = $1;
 206             last;
 207         }
 208     }
 209 }
 210
 211 =head2 _create_feature
 212
 213  Title   : _create_feature
 214  Usage   : $self->create_feature(\%feature)
 215  Function: Internal (not to be used directly)
 216  Returns : hash of feature values
 217  Args    : hash of more feature values
 218
 219 =cut
 220
 221 sub _create_feature {
 222     my ($self, $feat) = @_;
 223
 224     # create feature object
 225     my $feature = Bio::SeqFeature::Generic->new(
 226          -seq_id      => $feat->{name},
 227          -start       => $feat->{start},
 228          -end         => $feat->{end},
 229          -score       => $feat->{score},
 230          -source      => $feat->{source},
 231          -primary     => $feat->{primary},
 232          -logic_name  => $feat->{logic_name},
 233     );
 234
 235     $feature->score($feat->{peptideProb});
 236     $feature->add_tag_value('peptideProb', $feat->{peptideProb});
 237     $feature->add_tag_value('anchorProb', $feat->{anchorProb});
 238     $feature->add_tag_value('evalue',$feat->{anchorProb});
 239     $feature->add_tag_value('percent_id','NULL');
 240     $feature->add_tag_value("hid",$feat->{primary});
 241     $feature->add_tag_value('SignalpPrediction', $feat->{hmmProdiction});
 242     return $feature;
 243
 244 }
 245
 246 =head2 _seqname
 247
 248  Title   : _seqname
 249  Usage   : $self->_seqname($name)
 250  Function: Internal (not to be used directly)
 251  Returns :
 252  Args    :
 253
 254 =cut
 255
 256 sub _seqname{
 257     my ($self,$seqname)=@_;
 258
 259     if (defined$seqname){
 260         $self->{'seqname'}=$seqname;
 261     }
 262     return $self->{'seqname'};
 263 }
 264
 265 =head2 _fact1
 266
 267  Title   : _fact1
 268  Usage   : $self->fact1($fact1)
 269  Function: Internal (not to be used directly)
 270  Returns :
 271  Args    :
 272
 273 =cut
 274
 275 sub _fact1{
 276     my ($self, $fact1)=@_;
 277
 278     if (defined $fact1){
 279        $self->{'fact1'}=$fact1;
 280     }
 281     return $self->{'fact1'};
 282 }
 283
 284
 285
 286 1;
 287
 288