parallelize_plink_ibdibd.pl

parallelize_plink_ibd.pl

#!/usr/bin/perl
#
#Parses a .fam file and generates N number of sample list files depending on 2nd argument
#Generates PBS cluster submission files
#Justin Giles, Vanderbilt University, 2010
#
use strict;

if(scalar(@ARGV) != 3){
	print "\nPLEASE NOTE: EDIT THIS SCRIPT IN THE PBS AREA TO MODIFY IT TO YOUR NEEDS\n";
	print "parallelize_plink_ibd.pl <bedfile_base> <chunksize> <PLINK_frq_file>\n";
	print "\nExample:\n";
	print "parallelize_plink_ibd.pl /scratch/turnersd/dataprefix 100 /scratch/turnersd/freq.frq\n";
	exit;
}

my $bfile_base=$ARGV[0];
my $NUM_IN_FILE=$ARGV[1];
my $freq_file = $ARGV[2];
my $num_fam=`wc $bfile_base.fam | awk '{ print $1 }'`;
`gawk '{print \$1,\$2}' $bfile_base.fam | split -d -a 3 -l $NUM_IN_FILE - tmp.list`;

my $path = `pwd`;
chomp($path);

my $i=0;
my $a=`ls tmp* | wc | awk '{ print $1 }'`-1;
my $j=0;


while( $i <= $a ){
      while($j <= $a ){
		  my $file = "submitme_" . $i . "_" . $j . ".pbs";
		open(OUT, ">$file") || die "Cannot open $file\n";

	    print OUT "#!/bin/ksh\n
##EDIT THIS TO YOUR EMAIL##
#PBS -M username\@domain.com
#PBS -m bae

##CHANGE THIS IF YOU NEED MORE NODES FOR THIS PROCESS##
#PBS -l nodes=1

##CHANGE WALLTIME AND CPUT TO MATCH THE LENGTH OF THIS PROCESS##
#PBS -l walltime=10:00:00\n
#PBS -l cput=10:00:00\n

##CHANGE THE PMEM AND MEM TO MATCH THE MEMORY REQUIREMENTS OF THIS PROCESS##
#PBS -l pmem=7000mb\n
#PBS -l mem=7000mb\n
#PBS -j oe\n

## CHANGE THE GROUP TO MATCH YOUR GROUP##
#PBS -W group_list=YOUR_GROUP_NAME


dir=/scratch/\$USER/ibd/\n

[ -d \$dir ] || mkdir \$dir

cd \$dir/\n\n";

		my $i_file = `printf "%03i\n" $i`;
		my $j_file = `printf "%03i\n" $j`;
		chomp($i_file);
		chomp($j_file);
		print OUT "plink --bfile $bfile_base \\
 --read-freq $freq_file \\
 --genome --min 0.05 \\
 --genome-lists $path/tmp.list$i_file \\
				$path/tmp.list$j_file \\
 --out data.sub.$i.$j\n";
		close(OUT);
      	$j=$j+1;
	  }

       $i=$i+1;
     $j=$i;
 }