sa2sar

#!/usr/bin/perl

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

=head1 NAME

sa2sar

=head1 SYNOPSIS

 sa2sar [options] sa-file 

  -h     help
  saFile sa file
  		 a)sample-id
  		 b)father
  		 c)mother
 
 example: sa2sar hapmap.sar
         
 Extracts relative pairs from a sa-file.
       
=head1 DESCRIPTION

=cut

#option variables
my $help;
my $saFile;
my $sarFile;
my $colNo;
my %label2Column;

#data structures
my $headerProcessed;

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

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

$saFile = $ARGV[0];

my($name, $path, $ext) = fileparse($saFile, '\..*');
$sarFile = "$name.sar";
		  
#read sample annotation
open(SA, $saFile) || die "Cannot open $saFile";
open(SAR, ">$sarFile") || die "Cannot open $sarFile";
print SAR "sample-pair-id\tsample-id-1\tsample-id-2\trelationship\n";

$headerProcessed = 0;
while(<SA>)
{
    s/\r?\n?$//;
    
    if(!$headerProcessed)
    {
        $colNo = s/\t/\t/g + 1;
        
        my @fields = split('\t', $_, $colNo);
         
        SEARCH_LABEL: for my $label ('sample-id', 'father', 'mother')
        {
            for my $col (0 .. $#fields)
            {
                if ($fields[$col] eq $label)
                {           
                    $label2Column{$label}=$col;
                    next SEARCH_LABEL;
                }
            }
        }
        
        $headerProcessed = 1;
    }
    else
    {
        my @fields = split('\t', $_, $colNo);
        my $sampleID = $fields[$label2Column{'sample-id'}];
        my $fatherID = $fields[$label2Column{'father'}];
        my $motherID = $fields[$label2Column{'mother'}];
        
        if ($fatherID ne 'n/a')
        {
        	if (($sampleID cmp $fatherID) < 0)
        	{
        		print SAR "$sampleID/$fatherID\t$sampleID\t$fatherID\tPO\n";
        	}
        	else
        	{
        		print SAR "$fatherID/$sampleID\t$fatherID\t$sampleID\tPO\n";        		
        	}
        }

        if ($motherID ne 'n/a')
        {
        	if (($sampleID cmp $motherID) < 0)
        	{
        		print SAR "$sampleID/$motherID\t$sampleID\t$motherID\tPO\n";
        	}
        	else
        	{
        		print SAR "$motherID/$sampleID\t$motherID\t$sampleID\tPO\n";        		
        	}
        }
    }
}
close(SA);
close(SAR);