Bio::Tools GFF
SummaryIncluded librariesPackage variablesSynopsisDescriptionGeneral documentationMethods
Summary
Bio::Tools::GFF - A Bio::SeqAnalysisParserI compliant GFF format parser
Package variables
Privates (from "my" definitions)
$i = 0
$gff3_featureID = 0
%GFF3_ID_Tags = map { $_ => $i++ } qw(ID Parent Target)
Included modules
Bio::LocatableSeq
Bio::Seq::SeqFactory
Bio::SeqFeature::Generic
Inherit
Bio::Root::IO Bio::Root::Root Bio::SeqAnalysisParserI
Synopsis
    use Bio::Tools::GFF;
# specify input via -fh or -file my $gffio = Bio::Tools::GFF->new(-fh => \*STDIN, -gff_version => 2); my $feature; # loop over the input stream while($feature = $gffio->next_feature()) { # do something with feature } $gffio->close(); # you can also obtain a GFF parser as a SeqAnalasisParserI in # HT analysis pipelines (see Bio::SeqAnalysisParserI and # Bio::Factory::SeqAnalysisParserFactory) my $factory = Bio::Factory::SeqAnalysisParserFactory->new(); my $parser = $factory->get_parser(-input => \*STDIN, -method => "gff"); while($feature = $parser->next_feature()) { # do something with feature }
Description
This class provides a simple GFF parser and writer. In the sense of a
SeqAnalysisParser, it parses an input file or stream into SeqFeatureI
objects, but is not in any way specific to a particular analysis
program and the output that program produces.
That is, if you can get your analysis program spit out GFF, here is
your result parser.
Methods
_incrementGFF3ID
No description
Code
newDescriptionCode
_parse_headerDescriptionCode
_parse_sequence
No description
Code
next_segmentDescriptionCode
next_featureDescriptionCode
_feature_idx_by_seq_id
No description
Code
from_gff_stringDescriptionCode
_from_gff1_stringDescriptionCode
_from_gff2_stringDescriptionCode
_from_gff3_string
No description
Code
unescape
No description
Code
write_featureDescriptionCode
gff_stringDescriptionCode
_gff1_stringDescriptionCode
_gff2_stringDescriptionCode
_gff25_stringDescriptionCode
_gff3_stringDescriptionCode
gff_versionDescriptionCode
newFhDescriptionCode
fhDescriptionCode
_seq_by_id_h
No description
Code
get_seqsDescriptionCode
features_attached_to_seqsDescriptionCode
ignore_sequenceDescriptionCode
DESTROY
No description
Code
TIEHANDLE
No description
Code
READLINE
No description
Code
PRINT
No description
Code
Methods description
newcode    nextTop
 Title   : new
Usage : my $parser = Bio::Tools::GFF->new(-gff_version => 2,
-file => "filename.gff");
or
my $writer = Bio::Tools::GFF->new(-gff_version => 3,
-file => ">filename.gff3");
Function: Creates a new instance. Recognized named parameters are -file, -fh,
and -gff_version.
Returns : a new object
Args : named parameters
-gff_version => [1,2,3]
_parse_headercodeprevnextTop
 Title   : _parse_header
Usage : $gffio->_parse_header()
Function: used to turn parse GFF header lines. currently
produces Bio::LocatableSeq objects from ##sequence-region
lines
Returns : 1 on success
Args : none
next_segmentcodeprevnextTop
 Title   : next_segment
Usage : my $seq = $gffio->next_segment;
Function: Returns a Bio::LocatableSeq object corresponding to a
GFF "##sequence-region" header line.
Example :
Returns : A Bio::LocatableSeq object, or undef if
there are no more sequences.
Args : none
next_featurecodeprevnextTop
 Title   : next_feature
Usage : $seqfeature = $gffio->next_feature();
Function: Returns the next feature available in the input file or stream, or
undef if there are no more features.
Example :
Returns : A Bio::SeqFeatureI implementing object, or undef if there are no
more features.
Args : none
from_gff_stringcodeprevnextTop
 Title   : from_gff_string
Usage : $gff->from_gff_string($feature, $gff_string);
Function: Sets properties of a SeqFeatureI object from a GFF-formatted
string. Interpretation of the string depends on the version
that has been specified at initialization.
This method is used by next_feature(). It actually dispatches to one of the version-specific (private) methods. Example : Returns : void Args : A Bio::SeqFeatureI implementing object to be initialized The GFF-formatted string to initialize it from
_from_gff1_stringcodeprevnextTop
 Title   : _from_gff1_string
Usage :
Function:
Example :
Returns : void
Args : A Bio::SeqFeatureI implementing object to be initialized
The GFF-formatted string to initialize it from
_from_gff2_stringcodeprevnextTop
 Title   : _from_gff2_string
Usage :
Function:
Example :
Returns : void
Args : A Bio::SeqFeatureI implementing object to be initialized
The GFF2-formatted string to initialize it from
write_featurecodeprevnextTop
 Title   : write_feature
Usage : $gffio->write_feature($feature);
Function: Writes the specified SeqFeatureI object in GFF format to the stream
associated with this instance.
Returns : none
Args : An array of Bio::SeqFeatureI implementing objects to be serialized
gff_stringcodeprevnextTop
 Title   : gff_string
Usage : $gffstr = $gffio->gff_string($feature);
Function: Obtain the GFF-formatted representation of a SeqFeatureI object.
The formatting depends on the version specified at initialization.
This method is used by write_feature(). It actually dispatches to one of the version-specific (private) methods. Example : Returns : A GFF-formatted string representation of the SeqFeature Args : A Bio::SeqFeatureI implementing object to be GFF-stringified
_gff1_stringcodeprevnextTop
 Title   : _gff1_string
Usage : $gffstr = $gffio->_gff1_string
Function:
Example :
Returns : A GFF1-formatted string representation of the SeqFeature
Args : A Bio::SeqFeatureI implementing object to be GFF-stringified
_gff2_stringcodeprevnextTop
 Title   : _gff2_string
Usage : $gffstr = $gffio->_gff2_string
Function:
Example :
Returns : A GFF2-formatted string representation of the SeqFeature
Args : A Bio::SeqFeatureI implementing object to be GFF2-stringified
_gff25_stringcodeprevnextTop
 Title   : _gff25_string
Usage : $gffstr = $gffio->_gff2_string
Function: To get a format of GFF that is peculiar to Gbrowse/Bio::DB::GFF
Example :
Returns : A GFF2.5-formatted string representation of the SeqFeature
Args : A Bio::SeqFeatureI implementing object to be GFF2.5-stringified
_gff3_stringcodeprevnextTop
  Title   : _gff3_string
Usage : $gffstr = $gffio->_gff3_string
Function:
Example :
Returns : A GFF3-formatted string representation of the SeqFeature
Args : A Bio::SeqFeatureI implementing object to be GFF3-stringified
gff_versioncodeprevnextTop
  Title   : _gff_version
Usage : $gffversion = $gffio->gff_version
Function:
Example :
Returns : The GFF version this parser will accept and emit.
Args : none
newFhcodeprevnextTop
 Title   : newFh
Usage : $fh = Bio::Tools::GFF->newFh(-file=>$filename,-format=>'Format')
Function: does a new() followed by an fh()
Example : $fh = Bio::Tools::GFF->newFh(-file=>$filename,-format=>'Format')
$feature = <$fh>; # read a feature object
print $fh $feature; # write a feature object
Returns : filehandle tied to the Bio::Tools::GFF class
Args :
fhcodeprevnextTop
 Title   : fh
Usage : $obj->fh
Function:
Example : $fh = $obj->fh; # make a tied filehandle
$feature = <$fh>; # read a feature object
print $fh $feature; # write a feature object
Returns : filehandle tied to Bio::Tools::GFF class
Args : none
get_seqscodeprevnextTop
 Title   : get_seqs
Usage :
Function: Returns all Bio::Seq objects populated by GFF3 file
Example :
Returns :
Args :
features_attached_to_seqscodeprevnextTop
 Title   : features_attached_to_seqs
Usage : $obj->features_attached_to_seqs(1);
Function: For use with GFF3 containg sequence only
Setting this before parsing ensures that all Bio::Seq object
created will have the appropriate features added to them
defaults to false (off) Note that this mode will incur higher memory usage because features will have to be cached until the relevant feature comes along Example : Returns : value of features_attached_to_seqs (a boolean) Args : on set, new value (a boolean, optional)
ignore_sequencecodeprevnextTop
 Title   : ignore_sequence
Usage : $obj->ignore_sequence(1);
Function: For use with GFF3 containg sequence only
Setting this before parsing means that all sequence data will be
ignored
Example : Returns : value of ignore_sequence (a boolean) Args : on set, new value (a boolean, optional)
Methods code
_incrementGFF3IDdescriptionprevnextTop
sub _incrementGFF3ID {
        my ($self) = @_;
        return ++ $gff3_featureID;
    }
}
newdescriptionprevnextTop
sub new {
  my ($class, @args) = @_;
  my $self = $class->SUPER::new(@args);
  
  my ($gff_version, $noparse) = $self->_rearrange([qw(GFF_VERSION NOPARSE)],@args);

  # initialize IO
$self->_initialize_io(@args); $self->_parse_header() unless $noparse; $gff_version ||= 2; if( ! $self->gff_version($gff_version) ) { $self->throw("Can't build a GFF object with the unknown version ". $gff_version); } $self->{'_first'} = 1; return $self;
}
_parse_headerdescriptionprevnextTop
sub _parse_header {
   my ($self) = @_;

   my @unhandled;
   local $^W = 0; # hide warnings when we try and parse from a file opened
# for writing - there isn't really a better way to do
# AFAIK - cannot detech if a FH is read or write.
while(my $line = $self->_readline()){ my $handled = 0; next if /^\s+$/; if($line =~ /^\#\#sequence-region\s+(\S+)\s+(\S+)\s+(\S+)\s*/){ my($seqid,$start,$end) = ($1,$2,$3); push @{ $self->{'segments'} }, Bio::LocatableSeq->new ( -id => unescape($seqid), -start => $start, -end => $end, -length => ($end - $start + 1), ## make the length explicit
); $handled = 1; } elsif($line =~ /^(\#\#feature-ontology)/) { #to be implemented
$self->warn("$1 header tag parsing unimplemented"); } elsif($line =~ /^(\#\#attribute-ontology)/) { #to be implemented
$self->warn("$1 header tag parsing unimplemented"); } elsif($line =~ /^(\#\#source-ontology)/) { #to be implemented
$self->warn("$1 header tag parsing unimplemented"); } elsif($line =~ /^(\#\#\#)/) { #to be implemented
$self->warn("$1 header tag parsing unimplemented"); } elsif($line =~ /^(\#\#FASTA)/) { # initial ##FASTA is optional - artemis does not use it
$line = $self->_readline(); if ($line !~ /^\>(\S+)/) { $self->throw("##FASTA directive must be followed by fasta header, not: $line"); } } else { } if ($line =~ /^\>(.*)/) { # seq data can be at header or footer
my $seq = $self->_parse_sequence($line); if ($seq) { $self->_seq_by_id_h->{$seq->primary_id} = $seq; } } if(!$handled){ push @unhandled, $line } #looks like the header is over!
last unless $line =~ /^\#/; } foreach my $line (@unhandled){ $self->_pushback($line); } return 1;
}
_parse_sequencedescriptionprevnextTop
sub _parse_sequence {
    my ($self, $line) = @_;

    if ($line =~ /^\>(.*)/) {
        
        my $seqid = $1;
        $seqid =~ s/\s+$//;
        my $desc = '';
        if ($seqid =~ /(\S+)\s+(.*)/) {
            ($seqid, $desc) = ($1,$2);
        }
        my $res = '';
        while (my $line = $self->_readline) {
            if ($line =~ /^\#/) {
                last;
            }
            if ($line =~ /^\>/) {
                $self->_pushback($line);
                last;
            }
            $line =~ s/\s//g;
            $res .= $line;
        }
        return if $self->ignore_sequence;

        my $seqfactory = Bio::Seq::SeqFactory->new('Bio::Seq');
        my $seq = $seqfactory->create(-seq=>$res, 
                                      -id=>$seqid,
                                      -desc=>$desc);
        $seq->accession_number($seqid);
        if ($self->features_attached_to_seqs) {
            my @feats = 
              @{$self->_feature_idx_by_seq_id->{$seqid}};
            $seq->add_SeqFeature($_) foreach @feats;
            @{$self->_feature_idx_by_seq_id->{$seqid}} = ();
        }
        return $seq;
    }
    else {
        $self->throw("expected fasta header, not: $line");
    }
}
next_segmentdescriptionprevnextTop
sub next_segment {
   my ($self,@args) = @_;
   return shift @{ $self->{'segments'} } if defined $self->{'segments'};
   return;
}
next_featuredescriptionprevnextTop
sub next_feature {
    my ($self) = @_;
    
    my $gff_string;
    
    # be graceful about empty lines or comments, and make sure we return undef
# if the input's consumed
while(($gff_string = $self->_readline()) && defined($gff_string)) { if ($gff_string =~ /^\#\#\#/) { # all forward refs have been seen; TODO
} next if($gff_string =~ /^\#/ || $gff_string =~ /^\s*$/ || $gff_string =~ m{^//}); while ($gff_string =~ /^\>(.+)/) { # fasta can be in header or footer
my $seq = $self->_parse_sequence($gff_string); if ($seq) { $self->_seq_by_id_h->{$seq->primary_id} = $seq; $gff_string = $self->_readline; last unless $gff_string; } } last; } return unless $gff_string; my $feat = Bio::SeqFeature::Generic->new(); $self->from_gff_string($feat, $gff_string); if ($self->features_attached_to_seqs) { push(@{$self->_feature_idx_by_seq_id->{$feat->seq_id}}, $feat); } return $feat;
}
_feature_idx_by_seq_iddescriptionprevnextTop
sub _feature_idx_by_seq_id {
    my $self = shift;
    $self->{__feature_idx_by_seq_id} = shift if @_;
    $self->{__feature_idx_by_seq_id} = {}
      unless $self->{__feature_idx_by_seq_id};
    return $self->{__feature_idx_by_seq_id};
}
from_gff_stringdescriptionprevnextTop
sub from_gff_string {
    my ($self, $feat, $gff_string) = @_;

    if($self->gff_version() == 1)  {
	return $self->_from_gff1_string($feat, $gff_string);
    } elsif( $self->gff_version() == 3 ) {
	return $self->_from_gff3_string($feat, $gff_string);
    } else {
	return $self->_from_gff2_string($feat, $gff_string);
    }
}
_from_gff1_stringdescriptionprevnextTop
sub _from_gff1_string {
   my ($gff, $feat, $string) = @_;
   chomp $string;
   my ($seqname, $source, $primary, $start, $end, $score, 
       $strand, $frame, @group) = split(/\t/, $string);

   if ( !defined $frame ) {
       $feat->throw("[$string] does not look like GFF to me");
   }
   $frame = 0 unless( $frame =~ /^\d+$/);
   $feat->seq_id($seqname);
   $feat->source_tag($source);
   $feat->primary_tag($primary);
   $feat->start($start);
   $feat->end($end);
   $feat->frame($frame);
   if ( $score eq '.' ) {
       #$feat->score(undef);
} else { $feat->score($score); } if ( $strand eq '-' ) { $feat->strand(-1); } if ( $strand eq '+' ) { $feat->strand(1); } if ( $strand eq '.' ) { $feat->strand(0); } foreach my $g ( @group ) { if ( $g =~ /(\S+)=(\S+)/ ) { my $tag = $1; my $value = $2; $feat->add_tag_value($1, $2); } else { $feat->add_tag_value('group', $g); } }
}
_from_gff2_stringdescriptionprevnextTop
sub _from_gff2_string {
   my ($gff, $feat, $string) = @_;
   chomp($string);

   # according to the Sanger website, GFF2 should be single-tab
# separated elements, and the free-text at the end should contain
# text-translated tab symbols but no "real" tabs, so splitting on
# \t is safe, and $attribs gets the entire attributes field to be
# parsed later
# sendu: but the tag value pair can (should?) be separated by a tab. The
# 'no tabs' thing seems to apply only to the free text that is allowed for
# the value
my ($seqname, $source, $primary, $start, $end, $score, $strand, $frame, @attribs) = split(/\t+/, $string); my $attribs = join ' ', @attribs; if ( !defined $frame ) { $feat->throw("[$string] does not look like GFF2 to me"); } $feat->seq_id($seqname); $feat->source_tag($source); $feat->primary_tag($primary); $feat->start($start); $feat->end($end); $feat->frame($frame); if ( $score eq '.' ) { # $feat->score(undef);
} else { $feat->score($score); } if ( $strand eq '-' ) { $feat->strand(-1); } if ( $strand eq '+' ) { $feat->strand(1); } if ( $strand eq '.' ) { $feat->strand(0); } # <Begin Inefficient Code from Mark Wilkinson>
# this routine is necessay to allow the presence of semicolons in
# quoted text Semicolons are the delimiting character for new
# tag/value attributes. it is more or less a "state" machine, with
# the "quoted" flag going up and down as we pass thorugh quotes to
# distinguish free-text semicolon and hash symbols from GFF control
# characters
my $flag = 0; # this could be changed to a bit and just be twiddled
my @parsed; # run through each character one at a time and check it
# NOTE: changed to foreach loop which is more efficient in perl
# --jasons
for my $a ( split //, $attribs ) { # flag up on entering quoted text, down on leaving it
if( $a eq '"') { $flag = ( $flag == 0 ) ? 1:0 } elsif( $a eq ';' && $flag ) { $a = "INSERT_SEMICOLON_HERE"} elsif( $a eq '#' && ! $flag ) { last } push @parsed, $a; } $attribs = join "", @parsed; # rejoin into a single string
# <End Inefficient Code>
# Please feel free to fix this and make it more "perlish"
my @key_vals = split /;/, $attribs; # attributes are semicolon-delimited
foreach my $pair ( @key_vals ) { # replace semicolons that were removed from free-text above.
$pair =~ s/INSERT_SEMICOLON_HERE/;/g; # separate the key from the value
my ($blank, $key, $values) = split /^\s*([\w\d]+)\s/, $pair; if( defined $values ) { my @values; # free text is quoted, so match each free-text block
# and remove it from the $values string
while ($values =~ s/"(.*?)"//){ # and push it on to the list of values (tags may have
# more than one value... and the value may be undef)
push @values, $1; } # and what is left over should be space-separated
# non-free-text values
my @othervals = split /\s+/, $values; foreach my $othervalue(@othervals){ # get rid of any empty strings which might
# result from the split
if (CORE::length($othervalue) > 0) {push @values, $othervalue} } foreach my $value(@values){ $feat->add_tag_value($key, $value); } } }
}
_from_gff3_stringdescriptionprevnextTop
sub _from_gff3_string {
    my ($gff, $feat, $string) = @_;
    chomp($string);

    # according to the now nearly final GFF3 spec, columns should 
# be tab separated, allowing unescaped spaces to occur in
# column 9
my ($seqname, $source, $primary, $start, $end, $score, $strand, $frame, $groups) = split(/\t/, $string); if ( ! defined $frame ) { $feat->throw("[$string] does not look like GFF3 to me"); } $feat->seq_id($seqname); $feat->source_tag($source); $feat->primary_tag($primary); $feat->start($start); $feat->end($end); $feat->frame($frame); if ( $score eq '.' ) { #$feat->score(undef);
} else { $feat->score($score); } if ( $strand eq '-' ) { $feat->strand(-1); } if ( $strand eq '+' ) { $feat->strand(1); } if ( $strand eq '.' ) { $feat->strand(0); } my @groups = split(/\s*;\s*/, $groups); for my $group (@groups) { my ($tag,$value) = split /=/,$group; $tag = unescape($tag); my @values = map {unescape($_)} split /,/,$value; for my $v ( @values ) { $feat->add_tag_value($tag,$v); } } } # taken from Bio::DB::GFF
}
unescapedescriptionprevnextTop
sub unescape {
  my $v = shift;
  $v =~ tr/+/ /;
  $v =~ s/%([0-9a-fA-F]{2})/chr hex($1)/ge;
  return $v;
}
write_featuredescriptionprevnextTop
sub write_feature {
    my ($self, @features) = @_;
    return unless @features;
    if( $self->{'_first'} && $self->gff_version() == 3 ) {
	$self->_print("##gff-version 3\n");
    }
    $self->{'_first'} = 0;
    foreach my $feature ( @features ) {
	$self->_print($self->gff_string($feature)."\n");
    }
}
gff_stringdescriptionprevnextTop
sub gff_string {
    my ($self, $feature) = @_;

    if($self->gff_version() == 1) {
    	return $self->_gff1_string($feature);
    } elsif( $self->gff_version() == 3 ) {
    	return $self->_gff3_string($feature);
    } elsif( $self->gff_version() == 2.5 ) {
    	return $self->_gff25_string($feature);
    } else {
    	return $self->_gff2_string($feature);
    }
}
_gff1_stringdescriptionprevnextTop
sub _gff1_string {
   my ($gff, $feat) = @_;
   my ($str,$score,$frame,$name,$strand);

   if( $feat->can('score') ) {
       $score = $feat->score();
   }
   $score = '.' unless defined $score;

   if( $feat->can('frame') ) {
       $frame = $feat->frame();
   }
   $frame = '.' unless defined $frame;

   $strand = $feat->strand();
   if(! $strand) {
       $strand = ".";
   } elsif( $strand == 1 ) {
       $strand = '+';
   } elsif ( $feat->strand == -1 ) {
       $strand = '-';
   }
   
   if( $feat->can('seqname') ) {
       $name = $feat->seq_id();
       $name ||= 'SEQ';
   } else {
       $name = 'SEQ';
   }


   $str = join("\t",
                 $name,
		 $feat->source_tag(),
		 $feat->primary_tag(),
		 $feat->start(),
		 $feat->end(),
		 $score,
		 $strand,
		 $frame);

   foreach my $tag ( $feat->all_tags ) {
       foreach my $value ( $feat->each_tag_value($tag) ) {
	   $str .= " $tag=$value" if $value;
       }
   }


   return $str;
}
_gff2_stringdescriptionprevnextTop
sub _gff2_string {
   my ($gff, $origfeat) = @_;
    my $feat;
   if ($origfeat->isa('Bio::SeqFeature::FeaturePair')){
       $feat = $origfeat->feature2;
   } else {
       $feat = $origfeat;
   }
   my ($str1, $str2,$score,$frame,$name,$strand);

   if( $feat->can('score') ) {
       $score = $feat->score();
   }
   $score = '.' unless defined $score;

   if( $feat->can('frame') ) {
       $frame = $feat->frame();
   }
   $frame = '.' unless defined $frame;

   $strand = $feat->strand();
   if(! $strand) {
       $strand = ".";
   } elsif( $strand == 1 ) {
       $strand = '+';
   } elsif ( $feat->strand == -1 ) {
       $strand = '-';
   }

   if( $feat->can('seqname') ) {
       $name = $feat->seq_id();
       $name ||= 'SEQ';
   } else {
       $name = 'SEQ';
   }
   $str1 = join("\t",
                 $name,
		 $feat->source_tag(),
		 $feat->primary_tag(),
		 $feat->start(),
		 $feat->end(),
		 $score,
		 $strand,
		 $frame);
   # the routine below is the only modification I made to the original
# ->gff_string routine (above) as on November 17th, 2000, the
# Sanger webpage describing GFF2 format reads: "From version 2
# onwards, the attribute field must have a tag value structure
# following the syntax used within objects in a .ace file,
# flattened onto one line by semicolon separators. Tags must be
# standard identifiers ([A-Za-z][A-Za-z0-9_]*). Free text values
# must be quoted with double quotes".
# MW
my @all_tags = $feat->all_tags; my @group; if (@all_tags) { # only play this game if it is worth playing...
foreach my $tag ( @all_tags ) { my @v; foreach my $value ( $feat->each_tag_value($tag) ) { unless( defined $value && length($value) ) { $value = '""'; } elsif ($value =~ /[^A-Za-z0-9_]/){ $value =~ s/\t/\\t/g; # substitute tab and newline
# characters
$value =~ s/\n/\\n/g; # to their UNIX equivalents
$value = '"' . $value . '" '; } # if the value contains
# anything other than valid
# tag/value characters, then
# quote it
push @v, $value; # for this tag (allowed in GFF2 and .ace format)
} push @group, "$tag ".join(" ", @v); } } $str2 .= join(' ; ', @group); # Add Target information for Feature Pairs
if( ! $feat->has_tag('Target') && # This is a bad hack IMHO
! $feat->has_tag('Group') && $origfeat->isa('Bio::SeqFeature::FeaturePair') ) { $str2 = sprintf("Target %s %d %d", $origfeat->feature1->seq_id, ( $origfeat->feature1->strand < 0 ? ( $origfeat->feature1->end, $origfeat->feature1->start) : ( $origfeat->feature1->start, $origfeat->feature1->end) )) . ($str2?" ; ".$str2:""); # need to put Target information before other tag/value pairs - mw
} return $str1."\t".$str2;
}
_gff25_stringdescriptionprevnextTop
sub _gff25_string {
    my ($gff, $origfeat) = @_;
    my $feat;
    if ($origfeat->isa('Bio::SeqFeature::FeaturePair')){
	$feat = $origfeat->feature2;
    } else {
	$feat = $origfeat;
    }
    my ($str1, $str2,$score,$frame,$name,$strand);

    if( $feat->can('score') ) {
	$score = $feat->score();
    }
    $score = '.' unless defined $score;

    if( $feat->can('frame') ) {
	$frame = $feat->frame();
    }
    $frame = '.' unless defined $frame;

    $strand = $feat->strand();
    if(! $strand) {
	$strand = ".";
    } elsif( $strand == 1 ) {
	$strand = '+';
    } elsif ( $feat->strand == -1 ) {
	$strand = '-';
    }

    if( $feat->can('seqname') ) {
	$name = $feat->seq_id();
	$name ||= 'SEQ';
    } else {
	$name = 'SEQ';
    }
    $str1 = join("\t",
                 $name,
		 $feat->source_tag(),
		 $feat->primary_tag(),
		 $feat->start(),
		 $feat->end(),
		 $score,
		 $strand,
		 $frame);

    my @all_tags = $feat->all_tags;
    my @group; my @firstgroup;
    if (@all_tags) {   # only play this game if it is worth playing...
foreach my $tag ( @all_tags ) { my @v; foreach my $value ( $feat->each_tag_value($tag) ) { unless( defined $value && length($value) ) { $value = '""'; } elsif ($value =~ /[^A-Za-z0-9_]/){ $value =~ s/\t/\\t/g; # substitute tab and newline
# characters
$value =~ s/\n/\\n/g; # to their UNIX equivalents
$value = '"' . $value . '" '; } # if the value contains
# anything other than valid
# tag/value characters, then
# quote it
push @v, $value; # for this tag (allowed in GFF2 and .ace format)
} if (($tag eq 'Group') || ($tag eq 'Target')){ # hopefully we wont get both...
push @firstgroup, "$tag ".join(" ", @v); } else { push @group, "$tag ".join(" ", @v); } } } $str2 = join(' ; ', (@firstgroup, @group)); # Add Target information for Feature Pairs
if( ! $feat->has_tag('Target') && # This is a bad hack IMHO
! $feat->has_tag('Group') && $origfeat->isa('Bio::SeqFeature::FeaturePair') ) { $str2 = sprintf("Target %s ; tstart %d ; tend %d", $origfeat->feature1->seq_id, ( $origfeat->feature1->strand < 0 ? ( $origfeat->feature1->end, $origfeat->feature1->start) : ( $origfeat->feature1->start, $origfeat->feature1->end) )) . ($str2?" ; ".$str2:""); # need to put the target info before other tag/value pairs - mw
} return $str1 . "\t". $str2;
}
_gff3_stringdescriptionprevnextTop
sub _gff3_string {
    my ($gff, $origfeat) = @_;
    my $feat;
    if ($origfeat->isa('Bio::SeqFeature::FeaturePair')){
	$feat = $origfeat->feature2;
    } else {
	$feat = $origfeat;
    }

    my $ID = $gff->_incrementGFF3ID();

    my ($score,$frame,$name,$strand);

    if( $feat->can('score') ) {
	$score = $feat->score();
    }
    $score = '.' unless defined $score;

    if( $feat->can('frame') ) {
	$frame = $feat->frame();
    }
    $frame = '.' unless defined $frame;

    $strand = $feat->strand();

    if(! $strand) {
	$strand = ".";
    } elsif( $strand == 1 ) {
	$strand = '+';
    } elsif ( $feat->strand == -1 ) {
	$strand = '-';
    }

    if( $feat->can('seqname') ) {
	$name = $feat->seq_id();
	$name ||= 'SEQ';
    } else {
	$name = 'SEQ';
    }

    my @groups;

    # force leading ID and Parent tags
my @all_tags = grep { ! exists $GFF3_ID_Tags{$_} } $feat->all_tags; for my $t ( sort { $GFF3_ID_Tags{$b} <=> $GFF3_ID_Tags{$a} } keys %GFF3_ID_Tags ) { unshift @all_tags, $t if $feat->has_tag($t); } for my $tag ( @all_tags ) { # next if $tag eq 'Target';
if ($tag eq 'Target' && ! $origfeat->isa('Bio::SeqFeature::FeaturePair')){ # simple Target,start,stop
my($target_id, $b,$e,$strand) = $feat->get_tag_values($tag); next unless(defined($e) && defined($b) && $target_id); ($b,$e)= ($e,$b) if(defined $strand && $strand<0); $target_id =~ s/([\t\n\r%&\=;,])/sprintf("%%%X",ord($1))/ge; push @groups, sprintf("Target=%s %d %d", $target_id,$b,$e); next; } my $valuestr; # a string which will hold one or more values
# for this tag, with quoted free text and
# space-separated individual values.
my @v; for my $value ( $feat->each_tag_value($tag) ) { if( defined $value && length($value) ) { #$value =~ tr/ /+/; #spaces are allowed now
if ( ref $value eq 'Bio::Annotation::Comment') { $value = $value->text; } if ($value =~ /[^a-zA-Z0-9\,\;\=\.:\%\^\*\$\@\!\+\_\?\-]/) { $value =~ s/\t/\\t/g; # substitute tab and newline
# characters
$value =~ s/\n/\\n/g; # to their UNIX equivalents
# Unescaped quotes are not allowed in GFF3
# $value = '"' . $value . '"';
} $value =~ s/([\t\n\r%&\=;,])/sprintf("%%%X",ord($1))/ge; } else { # if it is completely empty,
# then just make empty double
# quotes
$value = '""'; } push @v, $value; } # can we figure out how to improve this?
$tag= lcfirst($tag) unless ($tag =~ /^(ID|Name|Alias|Parent|Gap|Target|Derives_from|Note|Dbxref|Ontology_term)$/); push @groups, "$tag=".join(",",@v); } # Add Target information for Feature Pairs
if( $feat->has_tag('Target') && ! $feat->has_tag('Group') && $origfeat->isa('Bio::SeqFeature::FeaturePair') ) { my $target_id = $origfeat->feature1->seq_id; $target_id =~ s/([\t\n\r%&\=;,])/sprintf("%%%X",ord($1))/ge; push @groups, sprintf("Target=%s %d %d", $target_id, ( $origfeat->feature1->strand < 0 ? ( $origfeat->feature1->end, $origfeat->feature1->start) : ( $origfeat->feature1->start, $origfeat->feature1->end) )); } # unshift @groups, "ID=autogenerated$ID" unless ($feat->has_tag('ID'));
if ( $feat->can('name') && defined($feat->name) ) { # such as might be for Bio::DB::SeqFeature
unshift @groups, 'Name=' . $feat->name; } my $gff_string = ""; if ($feat->location->isa("Bio::Location::SplitLocationI")) { my @locs = $feat->location->each_Location; foreach my $loc (@locs) { $gff_string .= join("\t", $name, $feat->source_tag() || '.', $feat->primary_tag(), $loc->start(), $loc->end(), $score, $strand, $frame, join(';', @groups)) . "\n"; } chop $gff_string; return $gff_string; } else { $gff_string = join("\t", $name, $feat->source_tag() || '.', $feat->primary_tag(), $feat->start(), $feat->end(), $score, $strand, $frame, join(';', @groups)); } return $gff_string;
}
gff_versiondescriptionprevnextTop
sub gff_version {
    my ($self, $value) = @_;
    if(defined $value && grep {$value == $_ } ( 1, 2, 2.5, 3)) {
	$self->{'GFF_VERSION'} = $value;
    }
    return $self->{'GFF_VERSION'};
}

# Make filehandles
}
newFhdescriptionprevnextTop
sub newFh {
    my $class = shift;
    return unless my $self = $class->new(@_);
    return $self->fh;
}
fhdescriptionprevnextTop
sub fh {
    my $self = shift;
    my $class = ref($self) || $self;
    my $s = Symbol::gensym;
    tie $$s,$class,$self;
    return $s;
}

# This accessor is used for accessing the Bio::Seq objects from a GFF3
# file; if the file you are using has no sequence data you can ignore
# this accessor
# This accessor returns a hash reference containing Bio::Seq objects,
# indexed by Bio::Seq->primary_id
}
_seq_by_id_hdescriptionprevnextTop
sub _seq_by_id_h {
    my $self = shift;

    return $self->{'_seq_by_id_h'} = shift if @_;
    $self->{'_seq_by_id_h'} = {}
    unless $self->{'_seq_by_id_h'};
    return $self->{'_seq_by_id_h'};
}
get_seqsdescriptionprevnextTop
sub get_seqs {
    my ($self,@args) = @_;
    return values %{$self->_seq_by_id_h};
}
features_attached_to_seqsdescriptionprevnextTop
sub features_attached_to_seqs {
    my $self = shift;

    return $self->{'_features_attached_to_seqs'} = shift if @_;
    return $self->{'_features_attached_to_seqs'};
}
ignore_sequencedescriptionprevnextTop
sub ignore_sequence {
    my $self = shift;

    return $self->{'_ignore_sequence'} = shift if @_;
    return $self->{'_ignore_sequence'};
}
DESTROYdescriptionprevnextTop
sub DESTROY {
    my $self = shift;
    $self->close();
}
TIEHANDLEdescriptionprevnextTop
sub TIEHANDLE {
    my ($class,$val) = @_;
    return bless {'gffio' => $val}, $class;
}
READLINEdescriptionprevnextTop
sub READLINE {
    my $self = shift;
    return $self->{'gffio'}->next_feature() unless wantarray;
    my (@list, $obj);
    push @list, $obj while $obj = $self->{'gffio'}->next_feature();
    return @list;
}
PRINTdescriptionprevnextTop
sub PRINT {
    my $self = shift;
    $self->{'gffio'}->write_feature(@_);
}

1;
}
General documentation
GFF3 AND SEQUENCE DATATop
GFF3 supports sequence data; see
http://www.sequenceontology.org/gff3.shtml
There are a number of ways to deal with this -
If you call
  $gffio->ignore_sequence(1)
prior to parsing the sequence data is ignored; this is useful if you
just want the features. It avoids the memory overhead in building and
caching sequences
Alternatively, you can call either
  $gffio->get_seqs()
Or
  $gffio->seq_id_by_h()
At the end of parsing to get either a list or hashref of Bio::Seq
objects (see the documentation for each of these methods)
Note that these objects will not have the features attached - you have
to do this yourself, OR call
  $gffio->features_attached_to_seqs(1)
PRIOR to parsing; this will ensure that the Seqs have the features
attached; ie you will then be able to call
  $seq->get_SeqFeatures();
And use Bio::SeqIO methods
Note that auto-attaching the features to seqs will incur a higher
memory overhead as the features must be cached until the sequence data
is found
TODOTop
Make a Bio::SeqIO class specifically for GFF3 with sequence data
FEEDBACKTop
Mailing ListsTop
User feedback is an integral part of the evolution of this and other
Bioperl modules. Send your comments and suggestions preferably to one
of the Bioperl mailing lists. Your participation is much appreciated.
  bioperl-l@bioperl.org                  - General discussion
http://bioperl.org/wiki/Mailing_lists - About the mailing lists
Support Top
Please direct usage questions or support issues to the mailing list:
bioperl-l@bioperl.org
rather than to the module maintainer directly. Many experienced and
reponsive experts will be able look at the problem and quickly
address it. Please include a thorough description of the problem
with code and data examples if at all possible.
Reporting BugsTop
Report bugs to the Bioperl bug tracking system to help us keep track
the bugs and their resolution. Bug reports can be submitted the web:
  http://bugzilla.open-bio.org/
AUTHOR - Matthew PocockTop
Email mrp-at-sanger.ac.uk
CONTRIBUTORS Top
Jason Stajich, jason-at-biperl-dot-org
Chris Mungall, cjm-at-fruitfly-dot-org
Steffen Grossmann [SG], grossman at molgen.mpg.de
Malcolm Cook, mec-at-stowers-institute.org
APPENDIXTop
The rest of the documentation details each of the object methods. Internal methods are usually preceded with a _