exemplar2tg

#!/usr/bin/perl

use warnings;
use strict;
use fralib;
use File::Basename;
use Getopt::Long;
use Pod::Usage;

=head1 NAME

exemplar2tg

=head1 SYNOPSIS

 exemplar2tg [options] <exemplar-file>
  -h            help
  exemplar-file comma separated value file
    
 example: exemplar2tg pscalare.csv
 
 No marker file is required, it is the user's onus to ensure the marker file generated is correct.
 Converts exemplar [blank],0,1,2,3 encoding to fratools -1,0,1,2 encoding.
       
=head1 DESCRIPTION

=cut

#option variables
my $help;
my $exemplarFile;
my $tgFile;
my $colNo;
my $headerProcessed;

#initialize options
Getopt::Long::Configure ('bundling');

if(!GetOptions ('h'=>\$help)
   || scalar(@ARGV)!=1)
{
    if ($help)
    {
        pod2usage(-verbose => 2);
    }
    else
    {
        pod2usage(1);
    }
}

$exemplarFile = $ARGV[0];

open(EXEMPLAR, "$exemplarFile") || die "Cannot open $exemplarFile\n";

if (!defined($tgFile))
{
	my ($name, $dir, $ext) = fileparse($exemplarFile, '\..*');
	$tgFile = "$name.tg";
}

open(TG, ">$tgFile") || die "Cannot open $tgFile\n";
$headerProcessed = 0;
while (<EXEMPLAR>)
{
    s/\r?\n?$//;
   
    if(!$headerProcessed)
    {
    	$colNo = s/,/,/g + 1;
    	
    	my $header = $_;
    	$header =~ s/,/\t/g;
    	
    	my @fields = split('\t', $header, 2);
        print TG "snp-id\t$fields[1]\n";
        
        $headerProcessed = 1;
    }
    else
    {
    	my @fields = split(',', $_, $colNo);
    
        my $snp = $fields[0];
		print TG "$snp";
		
		for my $col (1..$#fields)
		{
			my $genotype = $fields[$col];
				
			if ($genotype == 1)
			{
				print TG "\t0";
			}
			elsif ($genotype == 2)
			{
				print TG "\t1";
			}
			elsif ($genotype == 3)
			{
				print TG "\t2";
			}
			elsif (!defined($genotype))
			{
				print TG "\t-1";
			}
			elsif ($genotype == 0)
			{
				print TG "\t-1";
			}			
			else
			{
				die "Encoding not valid for $snp at column $col : $genotype";
			}

		}

        print TG "\n";
    }
}

close(TG);
close(EXEMPLAR);