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

View of /FigKernelScripts/generate_genome_feature_tables_for_sprout.pl

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.16 - (download) (as text) (annotate)
Mon Dec 5 18:56:37 2005 UTC (14 years, 3 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, rast_rel_2008_04_23, 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, caBIG-05Apr06-00, 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, caBIG-13Feb06-00, rast_rel_2009_03_26, mgrast_dev_10262011, rast_rel_2008_11_24, rast_rel_2008_08_07, HEAD
Changes since 1.15: +17 -0 lines
Add license words.

#
# Copyright (c) 2003-2006 University of Chicago and Fellowship
# for Interpretations of Genomes. All Rights Reserved.
#
# This file is part of the SEED Toolkit.
# 
# The SEED Toolkit is free software. You can redistribute
# it and/or modify it under the terms of the SEED Toolkit
# Public License. 
#
# You should have received a copy of the SEED Toolkit Public License
# along with this program; if not write to the University of Chicago
# at info@ci.uchicago.edu or the Fellowship for Interpretation of
# Genomes at veronika@thefig.info or download a copy from
# http://www.theseed.org/LICENSE.TXT.
#


$seq_chunk_sz = 8000;
$max_span_sz  = 4000;

use FIG;
my $fig = new FIG;


$usage = "usage: generate_genome_feature_tables_for_sprout GenomesFile RelationsDir";

(($genomesF  = shift @ARGV) &&
 ($relations = shift @ARGV)
)
    || die $usage;

(-d $relations) || die "$relations does not exist";

open(GENOME,">$relations/Genome") || die "could not open $relations/Genome";
open(HASCONTIG,">$relations/HasContig") || die "could not open $relation/HasContig";
open(SEQUENCE,">$relations/Sequence") || die "could not open $relation/Sequence";
open(CONTIG,">$relations/Contig") || die "could not open $relation/Contig";
open(ISMADEUPOF,">$relations/IsMadeUpOf") || die "could not open $relation/IsMadeUpOf";
open(ISLOCATEDIN,">$relations/IsLocatedIn") || die "could not open $relation/IsLocatedIn";
open(FEATURE,">$relations/Feature") || die "could not open $relations/Feature";
open(FEATUREALIAS,">$relations/FeatureAlias") || die "could not open $relations/FeatureAlias";
open(FEATURETRANSLATION,">$relations/FeatureTranslation") 
    || die "could not open $relations/FeatureTranslation";

@genomes = map { ($_ =~ /^(\d+\.\d+)/) ? $1 : () } `cat $genomesF`;
foreach $genome (@genomes)
{
    print STDERR "processing $genome\n";
    @tmp = `cat $FIG_Config::organisms/$genome/GENOME`;
    $line = $tmp[0];
    if ($line =~ /^(\S+)\s+(\S+)(\s+(\S(.*\S)?))?$/)
    {
        ($genus,$species) = ($1,$2);
        $extra = $4 ? $4 : "";
        $extra .= " [$genome]";
        print GENOME join("\t",($genome,1,$genus,$species,$extra,$fig->taxonomy_of($genome))),"\n";
    }
    else
    {
        print STDERR "failed: $line";
    }
    opendir(TMP,"$FIG_Config::organisms/$genome") || die "could not open $FIG_Config::organisms/$genome";
    my @contig_files = map { "$FIG_Config::organisms/$genome/$_" } grep { $_ =~ /^contigs\d*/ } readdir(TMP);
    closedir(TMP);
    
    my $contig_file;
    foreach $contig_file (@contig_files)
    {
        open(TMP,"<$contig_file")
            || die "could not open $contig_file";
        $/ = "\n>";
        while (defined($_ = <TMP>))
        {
            chomp;
            if ($_ =~ /^>?(\S+)[^\n]*\n(.*)/s)
            {
                $id  =  $1;
                $seq =  $2;
                $id  =~ s/:/-/g;
                print CONTIG "$genome:$id\n";
                print HASCONTIG join("\t",($genome,"$genome:$id")),"\n";

                $seq =~ s/\s//g;
                $ln = length($seq);
                for ($i=0; ($i < $ln); $i += $seq_chunk_sz)
                {
                    $chunk = substr($seq,$i,($i < ($ln - $seq_chunk_sz)) ? $seq_chunk_sz : ($ln - $i));
                    print SEQUENCE join("\t",("$genome:$id:$i","unknown",$chunk)),"\n";
                    print ISMADEUPOF join("\t",("$genome:$id","$genome:$id:$i",length($chunk),$i)),"\n";
                }
            }
        }
        close(TMP);
    }
#   print STDERR "processing features\n";
    $/ = "\n";

    opendir(TMP,"$FIG_Config::organisms/$genome/Features") || die "could not open $FIG_Config::organisms/$genome/Features";
    @types = grep { ($_ =~ /^([a-zA-Z]\S*)/) && (-s "$FIG_Config::organisms/$genome/Features/$1/tbl") } readdir(TMP);
    closedir(TMP);

    foreach $type (@types)
    {
#       print STDERR "        $type\n";
        foreach $fid ($fig->all_features($genome,$type))
        {
            my $loc = $fig->feature_location($fid);
            my $tran;
            if (($type eq "peg") && ($tran = $fig->get_translation($fid)))
            {
                print FEATURETRANSLATION "$fid\t$tran\n";
            }
            print FEATURE join("\t",($fid,1,$type)),"\n";
            @loc = &split_long($max_span_sz,split(/,/,$loc));
            for ($i=0; ($i < @loc); $i++)
            {
                $span = $loc[$i];
                if ($span =~ /^(\S+)_(\d+)_(\d+)$/)
                {
                    ($contig,$beg,$end) = ($1,$2,$3);
                    if (($i == $#loc) && ($type eq "peg"))
                    {
                        if ($beg < $end)
                        {
                            $poss_stop = uc $fig->dna_seq($genome,join("_",($contig,$end+1,$end+3)));
                            if ($poss_stop =~ /(TAA|TGA|TAG)/)
                            {
                                $end += 3;
                            }
                        }
                        else
                        {
                            $poss_stop = uc $fig->dna_seq($genome,join("_",($contig,$end-1,$end-3)));
                            if ($poss_stop =~ /(TAA|TGA|TAG)/)
                            {
                                $end -= 3;
                            }
                        }
                    }
                    print ISLOCATEDIN join("\t",($fid,"$genome:$contig",$beg,($beg <= $end) ? "+" : "-",abs($end-$beg)+1,$i)),"\n";
                }
                else
                {
                    print STDERR "BAD LOCATION: $genome $fid $loc\n";
                }
            }
            my @aliases = grep { $_ =~ /^(([NXYZA]P_[0-9\.]+)|(gi\|\d+)|(kegg\|\S+)|(uni\|[A-Z0-9]{6})|(sp\|[A-Z0-9]{6})|([a-zA-Z]{3,4}\d?)|([a-zA-Z]{1,2}\d+))$/ }
                          $fig->feature_aliases($fid);
            
            foreach $_ (@aliases)
            {
                print FEATUREALIAS join("\t",($fid,$_)),"\n";
            }
        }
        close(TMP);
    }
}
close(OLD);
$/ = "\n";

close(GENOME);
close(COMESFROM);
print STDERR "Genome relation finished\n";

close(FEATURE);
close(FEATUREALIAS);
close(ISLOCATEDIN);
print STDERR "Feature relation finished\n";

close(CONTIG);
close(ISMADEUPOF);
close(SEQUENCE);
close(HASCONTIG);

print STDERR "Contig relation finished\n";

sub split_long {
    my($max_span_sz,@locs) = @_;
    my($contig,$beg,$end,$end1,$loc1);

    my(@locs2) = ();

    foreach $loc1 (@locs)
    {
        if ($loc1 =~ /^(\S+)_(\d+)_(\d+)$/)
        {
            ($contig,$beg,$end) = ($1,$2,$3);
            if (abs($beg - $end) <= $max_span_sz)
            { 
                push(@locs2,$loc1);
            }
            elsif ($beg < $end)
            {
                while ($beg < $end)
                {
                    $end1 = ($end < ($beg + $max_span_sz)) ? $end : ($beg + ($max_span_sz - 1));
                    push(@locs2,join("_",($contig,$beg,$end1)));
                    $beg += $max_span_sz;
                }
            }
            elsif ($beg > $end)
            {
                while ($beg > $end)
                {
                    $end1 = ($end > ($beg - $max_span_sz)) ? $end : ($beg - ($max_span_sz - 1));
                    push(@locs2,join("_",($contig,$beg,$end1)));
                    $beg -= $max_span_sz;
                }
            }
        }
        else
        {
            confess "bad location: $loc1";
        }
    }
    return @locs2;
}

MCS Webmaster
ViewVC Help
Powered by ViewVC 1.0.3