# TFBS module for TFBS::PatternGen::AnnSpec # # Copyright Wynand Alkema # # You may distribute this module under the same terms as perl itself # # POD =head1 NAME TFBS::PatternGen::AnnSpec - a pattern factory that uses the AnnSpec program =head1 SYNOPSIS my $patterngen = TFBS::PatternGen::AnnSpec->new(-seq_file=>'sequences.fa', -binary => 'ann-spec ' my $pfm = $patterngen->pattern(); # $pfm is now a TFBS::Matrix::PFM object =head1 DESCRIPTION TFBS::PatternGen::AnnSpec builds position frequency matrices using an external program AnnSpec (Workman, C. and Stormo, G.D. (2000) ANN-Spec: A method for discovering transcription factor binding sites with improved specificity. Proc. Pacific Symposium on Biocomputing 2000). =head1 FEEDBACK Please send bug reports and other comments to the author. =head1 AUTHOR - Wynand Alkema Wynand Alkema EWynand.Alkema@cgb.ki.seE =cut package TFBS::PatternGen::AnnSpec; use vars qw(@ISA); use strict; # Object preamble - inherits from TFBS::PatternGen; use TFBS::PatternGen; use TFBS::PatternGen::AnnSpec::Motif; use File::Temp qw(:POSIX); use Bio::Seq; use Bio::SeqIO; @ISA = qw(TFBS::PatternGen); =head2 new Title : new Usage : my $pattrengen = TFBS::PatternGen::AnnSpec->new(%args); Function: the constructor for the TFBS::PatternGen::AnnSpec object Returns : a TFBS::PatternGen::AnnSpec object Args : This method takes named arguments; you must specify one of the following three -seq_list # a reference to an array of strings # and/or Bio::Seq objects # or -seq_stream # A Bio::SeqIO object # or -seq_file # the name of the fasta file containing # all the sequences Other arguments are: -binary # a fully qualified path to the 'meme' executable # OPTIONAL: default 'ann-spec' -additional_params # a string containing additional # command-line switches for the # ann-spec program =cut sub new { my ($caller, %args) = @_; my $self = bless {}, ref($caller) || $caller; $self->{'filename'} =$args{'-seq_file'}; $self->{'additional_params'} = ($args{'-additional_params'} ? (ref($args{'-additional_params'}) ? join(' ', @{$args{'-additional_params'}}) : $args{'-additional_params'}) : "" ); $self->{'binary'} = $args{'-binary'} || 'ann-spec'; $self->_create_seq_set(%args) or die ('Error creating sequence set'); $self->_run_AnnSpec() or $self->throw("Error running AnnSpec."); return $self; } =head2 pattern =head2 all_patterns =head2 patternSet The three methods listed above are used for the retrieval of patterns, and are common to all TFBS::PatternGen::* classes. Please see L for details. =cut sub _run_AnnSpec{ my ($self)=shift; my $tmp_file = tmpnam(); my $outstream = Bio::SeqIO->new(-file=>">$tmp_file", -format=>"fasta"); foreach my $seqobj (@{ $self->{'seq_set'} } ) { $outstream->write_seq($seqobj); } $outstream->close(); my $command_line = $self->{'binary'}." ". "-p ".$tmp_file." ". # $self->{'motif_length_string'}." ". # $self->{'nr_hits_string'}." ". $self->{'additional_params'}. ""; # print STDERR "$command_line\n"; my $resultstring = `$command_line`; # print STDERR $resultstring; $self->_parse_AnnSpec_output($resultstring,$command_line); unlink $tmp_file; return 1 } sub _parse_AnnSpec_output{ my ($self,$resultstring,$command_line)=@_; if ($resultstring eq''){ # warn "Error running AnnSpec\nNo patterns produced"; $self->throw ("Error running AnnSpec using command:\n $command_line"); return; } my ($consensus,$matrix)=$self->_parse_raw_matrix($resultstring); my ($score,$sites)=$self->_parse_sites($resultstring); my $motif =TFBS::PatternGen::AnnSpec::Motif->new ( #-length => $length."", # -bg_probabilities => [split /\s+/, $raw_bp], -tags => {consensus => $consensus, score=>$score}, -nr_hits => 1, -sites=>$sites, -matrix => $matrix ); push @{ $self->{'motifs'} }, $motif; return } sub _parse_sites{ my ($self,$string)=@_; # print $raw_motif; my @hits; my ($sites)=$string=~/STR BEST_SITES\n(.*)STR ave\(S\)/s; my ($average)=$string=~/STR ave\(S\)\s+(\d*\.*\d*)/; my ($score)=$string=~/STR ln\(ave\(sum\(exp\(S\)\)\)\)\s+(\d*\.*\d*)/; # print STDERR $score,"\n"; my @sites=split/\n/,$sites; shift @sites; # print "@sites\n"; foreach my $site (@sites){ my @site_array=split(/\s+/,$site); # print "$site_array[3]\n"; # print "$site_array[5]\n"; my ($seq_id)=$site_array[5]=~/>(.*)/; my $strand=1; $strand=-1 if $site_array[3]=~/\'/;#MEans we have a pattern in the reverse strand my ($start)=$site_array[3]=~/(\d+)/; my $site = Bio::SeqFeature::Generic->new ( -start => $start, -end => $start+(length$site_array[4])-1, -strand => $strand, -source => 'AnnSpec', -score => $site_array[2], ); # foreach my $seq(@{$self->{'seq_set'}}){ if ($seq->id eq $seq_id){ $site->attach_seq ($seq); } } push (@hits,$site); } return $score,\@hits; } sub _parse_raw_matrix{ my ($self,$string)=@_; my ($matrix)=$string=~/ALR ALIGNMENT_MATRIX.*ALR\s+-+(.*)ALR CONSENSUS/s; my ($consensus)=$string=~/ALR CONSENSUS (.*)\n/; #print $consensus; my @matrix=split("\n",$matrix); shift @matrix; my @pfm; foreach my $row(@matrix){ # print $row; my @row=split /\s+/, $row; push @pfm, [@row[2..scalar@row-1]]; } return $consensus, \@pfm; } 1;