#!/usr/bin/perl
use strict;
use warnings;
use v5.14;

# Transpose columns to rows, in preparation for using BioPerl:

die "need two arguments (i.e. chr cont) at invocation" unless @ARGV == 2;
chomp( my $chr_num = shift );
chomp( my $cont    = shift );

open my $in_file, '<', "chr${chr_num}_exome_snps_only.vcf_$cont"
  or die "Cannot open input file: $!";

open my $out_file, '>', "chr${chr_num}_exome_snps_only_transposed_$cont"
  or die "Cannot open output file: $!";

my @rows       = ();
my @transposed = ();

while (<$in_file>) {
    chomp;
    push @rows, [split];
}

foreach my $row (@rows) {
    foreach my $column ( 0 .. $#{$row} ) {
        push( @{ $transposed[$column] }, $row->[$column] );
    }
}

foreach my $new_row (@transposed) {
    foreach my $new_col ( @{$new_row} ) {
        print $out_file $new_col, "\t";
    }
    print $out_file "\n";
}

# Remove unnecessary rows, clean up the SNPs and put into CSV format:

open my $new_in_file, '<', "chr${chr_num}_exome_snps_only_transposed_$cont"
  or die "Cannot open new input file: $!";
open my $new_out_file, '>', "chr${chr_num}_exome_snps_processed_$cont"
  or die "Cannot open new output file: $!";

while (<$new_in_file>) {
    chomp;
    next          if $. == 1;   # remove header line
    s/POS/SAMPLE/ if $. == 2;
    s/\s+/,/g     if $. == 2;   # replace whitespace with a comma on ID row
    next          if $. >= 3 and $. <= 9;   # remove unwanted lines
    s/\s+(\d)\|(\d)\S+/,$1 $2/g;    # clean up the SNPs & put into CSV format
    print $new_out_file "$_\n";
}



=cut

sample $in_file (note: hundreds more columns i.e. individuals beginning with HG, and hundreds of thousands more rows):

#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	HG00553	HG00554	HG00637
22	16287226	.	C	T	100	PASS	SNP;BAVGPOST=0.986;BRSQ=0.635;LDAF=0.0167;AVGPOST=0.9859;RSQ=0.6343;ERATE=0.0026;THETA=0.0022;AC=23;AN=2184	GT:DS:GL:BD	0|0:0.000:-0.48,-0.48,-0.48:0.021	0|0:0.000:-0.00,-2.69,-5.00:0.0001	0|0:0.000:-0.00,-2.53,-5.00:0.0001
22	16287365	.	C	T	100	PASS	SNP;BAVGPOST=0.998;BRSQ=0.540;LDAF=0.0030;AVGPOST=0.9959;RSQ=0.4172;ERATE=0.0009;THETA=0.0201;AC=2;AN=2184	GT:DS:GL:BD	0|0:0.000:-0.48,-0.48,-0.48:0.0024	0|0:0.000:-0.00,-4.40,-5.00:0	0|0:0.000:-0.02,-1.38,-5.00:0.0002
22	16287649	.	G	A	100	PASS	SNP;BAVGPOST=0.949;BRSQ=0.519;LDAF=0.0645;AVGPOST=0.9342;RSQ=0.5421;ERATE=0.0115;THETA=0.0095;AC=81;AN=2184	GT:DS:GL:BD	0|0:0.050:-0.17,-0.50,-2.67:0.072	0|0:0.000:-0.00,-2.34,-5.00:0.0002	0|0:0.000:-0.01,-1.86,-5.00:0.0024
22	16287784	.	C	T	100	PASS	SNP;BAVGPOST=0.998;BRSQ=0.844;LDAF=0.0089;AVGPOST=0.9941;RSQ=0.7151;ERATE=0.0030;THETA=0.0043;AC=13;AN=2184	GT:DS:GL:BD	0|0:0.000:-0.48,-0.48,-0.48:0.0104	0|0:0.000:-0.00,-2.79,-5.00:0.0001	0|0:0.000:0.00,-5.00,-5.00:0
22	16287851	.	G	A	100	PASS	SNP;BAVGPOST=0.939;BRSQ=0.764;LDAF=0.1473;AVGPOST=0.9183;RSQ=0.7385;ERATE=0.0015;THETA=0.0309;AC=280;AN=2184	GT:DS:GL:BD	0|0:0.100:-0.08,-0.78,-5.00:0.0406	0|0:0.000:-0.00,-2.76,-5.00:0	0|0:0.000:-0.00,-3.49,-5.00:0
22	16287912	.	G	A	100	PASS	SNP;BAVGPOST=0.998;BRSQ=0.800;LDAF=0.0057;AVGPOST=0.9983;RSQ=0.8580;ERATE=0.0006;THETA=0.0210;AC=11;AN=2184	GT:DS:GL:BD	0|0:0.000:-0.12,-0.63,-4.70:0.0028	0|0:0.000:-0.01,-1.91,-5.00:0.0003	0|0:0.000:-0.00,-2.30,-5.00:0


sample $new_out_file (note: hundreds of thousands more columns, hundreds more rows) 

SAMPLE,16287215,16287226,16287365,16287649,16287784
HG00553,0 0,0 0,0 0,0 0,0 0
HG00554,0 0,0 0,0 0,0 0,0 0
HG00637,0 0,0 0,0 0,0 0,0 0
HG00638,0 0,0 0,0 0,0 0,0 0
HG00640,0 0,0 0,0 0,0 0,0 0
HG00641,0 0,0 0,0 0,0 0,0 0
HG00731,0 0,0 0,0 0,0 0,0 0