[Bio] / FigKernelScripts / gff2ncbitab.pl Repository:
ViewVC logotype

View of /FigKernelScripts/gff2ncbitab.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.3 - (download) (as text) (annotate)
Fri May 23 21:07:21 2008 UTC (11 years, 9 months ago) by olson
Branch: MAIN
CVS Tags: mgrast_dev_08112011, rast_rel_2009_05_18, mgrast_dev_08022011, rast_rel_2014_0912, rast_rel_2008_06_18, myrast_rel40, rast_rel_2008_06_16, mgrast_dev_05262011, rast_rel_2008_12_18, mgrast_dev_04082011, rast_rel_2008_07_21, rast_rel_2010_0928, rast_2008_0924, mgrast_version_3_2, mgrast_dev_12152011, mgrast_dev_06072011, rast_rel_2008_09_30, rast_rel_2009_0925, rast_rel_2010_0526, rast_rel_2014_0729, mgrast_dev_02212011, rast_rel_2010_1206, mgrast_release_3_0, mgrast_dev_03252011, rast_rel_2010_0118, mgrast_rel_2008_0924, mgrast_rel_2008_1110_v2, rast_rel_2009_02_05, rast_rel_2011_0119, mgrast_rel_2008_0625, mgrast_release_3_0_4, mgrast_release_3_0_2, mgrast_release_3_0_3, mgrast_release_3_0_1, mgrast_dev_03312011, mgrast_release_3_1_2, mgrast_release_3_1_1, mgrast_release_3_1_0, mgrast_dev_04132011, rast_rel_2008_10_09, mgrast_dev_04012011, rast_release_2008_09_29, mgrast_rel_2008_0806, mgrast_rel_2008_0923, mgrast_rel_2008_0919, rast_rel_2009_07_09, rast_rel_2010_0827, mgrast_rel_2008_1110, myrast_33, rast_rel_2011_0928, rast_rel_2008_09_29, mgrast_rel_2008_0917, rast_rel_2008_10_29, mgrast_dev_04052011, mgrast_dev_02222011, rast_rel_2009_03_26, mgrast_dev_10262011, rast_rel_2008_11_24, rast_rel_2008_08_07, HEAD
Changes since 1.2: +3 -0 lines
add swissprot

use FIG_Config;
use FIG;
use Data::Dumper;

use strict;

my $usage = "$0 gff-file > tab-sep-data";
@ARGV == 1 or die $usage;

my $ffs = new FFs("$FIG_Config::data/FigfamsData");
$ffs or die "bad ffs\n";

my @desired_dbxref = qw(RefSeq_Prot
			IMG
			TIGR
			SP
			UniProtKB
			KEGG
			InterPro
			PFAM
			SMART
			Swiss-Prot
			);
my %desired_dbxref = map { $_ => 1} @desired_dbxref;

my $gff = shift;

open(GFF, "<", $gff) or die "cannot open gff file $gff: $!\n";

my %dat;
my $contig;
while (<GFF>)
{
    last if /^##FASTA/;
    next if /^\s*\#/;

    chomp;
    
    my($contig_id, undef, $what, $start, $end, undef, $strand, undef, $dat) = split(/\t/);

    my @parts = split(/;/, $dat);
    my %attr;
    for my $p (@parts)
    {
	if ($p =~ /^([^=]+)=(.*)/)
	{
	    $attr{$1} = [split(/,/, $2)];
	}
    }

    my @dbxref = map { [ split(/:/) ] } @{$attr{Dbxref}};

    print STDERR "$_\n" for map { $_->[0] } @dbxref;

    my $id = $attr{ID}->[0];

    if ($what eq 'contig')
    {
	$contig = $attr{Name}->[0];
	$contig =~ s/^Contig\s+//;
    }
    elsif ($what eq 'cds')
    {
	#
	# generate lines with
	# id contig coords refseq-prot evcode gene-sym descr dbxref kegg subsys weblink struct-desc proteinseq
	#
	my $fig_id = (map { $_->[1] } grep { $_->[0] eq 'FIG_ID' } @dbxref)[0];
	my $coords = $strand eq '+' ? "${start}_$end"  : "${end}_$start";
	my $refseq = (map { $_->[1] } grep { $_->[0] eq 'RefSeq_Prot' } @dbxref)[0];
	my $figfam = $ffs->family_containing_peg($fig_id);

	my $evcode = join(",", @{$attr{evidence_code}}) if $attr{evidence_code};

	my $gene_sym = $attr{gene_symbol}->[0] if $attr{gene_symbol};
	my $descr = $attr{description}->[0] if $attr{description};
	my $dbxref = join(",", map { join(":", @$_) } grep { $desired_dbxref{$_->[0]} } @dbxref);
	my $kegg = join(",", map { join(":", @$_) } grep { $_->[0] =~ /^KEGG/} @dbxref);
	my $weblink;
	if ($attr{web_id})
	{
	    $weblink = "http://www.nmpdr.org/linkin.cgi?id=$attr{web_id}->[0]";
	}
	$dat{$id} = [$fig_id, $contig, $coords, $refseq, $evcode, $gene_sym, $descr, $dbxref, $kegg, '', $weblink, $figfam];
    }
}

my $cur_id;
my $cur_seq;
while (<GFF>)
{
    chomp;
    if (/^>(\S+)/)
    {
	if ($cur_id and $dat{$cur_id})
	{
	    print join("\t", @{$dat{$cur_id}}, $cur_seq), "\n";
	}
	$cur_seq = '';
	$cur_id = $1;
    }
    else
    {
	s/\s+//g;
	$cur_seq .= $_;
    }
}
if ($cur_id and $dat{$cur_id})
{
    print join("\t", @{$dat{$cur_id}}, $cur_seq), "\n";
}



    

MCS Webmaster
ViewVC Help
Powered by ViewVC 1.0.3