branch-1-6/Bio/Matrix/PSM/IO/psiblast.pm

   1 #---------------------------------------------------------
   2 # $Id$
   3
   4 =head1 NAME
   5
   6 Bio::Matrix::PSM::IO::psiblast - PSM psiblast parser
   7
   8 =head1 SYNOPSIS
   9
  10 See Bio::Matrix::PSM::IO for documentation
  11
  12 =head1 DESCRIPTION
  13
  14 Parser for ASCII matrices from PSI-BLAST (blastpgp program in
  15 BLAST distribution).
  16
  17 =head1 FEEDBACK
  18
  19 =head2 Mailing Lists
  20
  21 User feedback is an integral part of the evolution of this and other
  22 Bioperl modules. Send your comments and suggestions preferably to one
  23 of the Bioperl mailing lists.  Your participation is much appreciated.
  24
  25   bioperl-l@bioperl.org                  - General discussion
  26   http://bioperl.org/wiki/Mailing_lists  - About the mailing lists
  27
  28 =head2 Support
  29
  30 Please direct usage questions or support issues to the mailing list:
  31
  32 I<bioperl-l@bioperl.org>
  33
  34 rather than to the module maintainer directly. Many experienced and
  35 reponsive experts will be able look at the problem and quickly
  36 address it. Please include a thorough description of the problem
  37 with code and data examples if at all possible.
  38
  39 =head2 Reporting Bugs
  40
  41 Report bugs to the Bioperl bug tracking system to help us keep track
  42 the bugs and their resolution.  Bug reports can be submitted via the
  43 web:
  44
  45   http://bugzilla.open-bio.org/
  46
  47 =head1 AUTHOR - James Thompson
  48
  49 Email tex@biosysadmin.com
  50
  51 =head1 APPENDIX
  52
  53 =cut
  54
  55
  56 # Let the code begin...
  57 package Bio::Matrix::PSM::IO::psiblast;
  58 use Bio::Matrix::PSM::Psm;
  59 use Bio::Matrix::PSM::ProtMatrix;
  60 use strict;
  61
  62 use base qw(Bio::Matrix::PSM::PsmHeader Bio::Matrix::PSM::IO);
  63
  64 # define the order in which amino acids are listed in the psiblast matrix file
  65 our @ordered_alphabet = qw/A  R  N  D  C  Q  E  G  H  I  L  K  M  F  P  S  T  W  Y  V/;
  66
  67 =head2 new
  68
  69  Title   : new
  70  Usage   : my $psmIO  =  Bio::Matrix::PSM::IO->new(-format=>'psiblast',
  71                                                   -file=>$file);
  72  Function: Associates a file with the appropriate parser
  73  Throws  :
  74  Example :
  75  Args    :
  76  Returns : Bio::Matrix::PSM::ProtMatrix->new(@args);
  77
  78 =cut
  79
  80 sub new {
  81    my ($class,@args)=@_;
  82    my $line;
  83
  84    my $self   = $class->SUPER::new(@args);
  85    my ($file) = $self->_rearrange(['FILE'], @args);
  86    $self->_initialize_io(@args) || warn "Did you intend to use STDIN?"; # Read only for now
  87    $self->_initialize;
  88
  89    $self->{_ordered_alphabet} = \@ordered_alphabet;
  90    return $self;
  91 }
  92
  93 =head2 next_psm
  94
  95  Title   : next_psm
  96  Usage   : my $psm = $psmIO->next_psm();
  97  Function: Reads the next PSM from the input file, associated with this object
  98  Throws  : None
  99  Returns : Bio::Matrix::PSM::ProtPsm object
 100  Args    : none
 101
 102 =cut
 103
 104 sub next_psm {
 105    my $self = shift;
 106    my $line;
 107
 108    return if ($self->{_end});
 109
 110    my %args;
 111    my @ordered_alphabet = @{$self->{_ordered_alphabet}};
 112
 113    while ( defined( $line = $self->_readline) ) {
 114       # remove leading and trailing whitespace
 115       chomp $line;
 116       $line =~ s/^\s+//g;
 117       $line =~ s/\s+$//g;
 118
 119       if ( $line =~ /^(\d+)\s+(\w{1})/ ) {  # match reference aa and position number
 120          my @elements = split /\s+/, $line;
 121
 122          my $position = shift @elements;
 123          my $letter   = shift @elements;
 124
 125          my $ratio = pop @elements;
 126          my $ic    = pop @elements;
 127
 128          # put the next 20 elements into the correct array in %args
 129          for ( 0 .. 19 ) { push @{$args{'l'.$ordered_alphabet[$_]}}, shift @elements; }
 130          for ( 0 .. 19 ) { push @{$args{'p'.$ordered_alphabet[$_]}}, shift @elements; }
 131
 132          push @{$args{'ic'}}, $ic;
 133       }
 134    }
 135
 136    $self->{_end} = 1; # psiblast matrix files currently only hold one PSM per file
 137
 138    my $psm = Bio::Matrix::PSM::ProtMatrix->new( %args );
 139    return $psm;
 140 }
 141
 142 sub DESTROY {
 143    my $self=shift;
 144    $self->close;
 145 }
 146
 147 1;