Hello ic23oluk,

I think you are looking for something like that (sample of code below):

#!/usr/bin/perl
use strict;
use warnings;
use Data::Dumper;

# CHECK FOR CORRECT USAGE
unless (@ARGV == 1){
    die "Usage: perl $0 \"file.txt\"\n";
}

my %hash;
while (<>) {
    chomp;
    next if /^\s*$/; # skip empty lines
    my @columns = split (/\t/, $_);
    next if $columns[0] =~  m/[^0-9.]/; # skipe lines that do no start
+ with a number
    $hash{$columns[1]}{$columns[2]} = $columns[4];
}

print Dumper \%hash;

__END__

$ perl test.pl file.txt
$VAR1 = {
          'Sample 1' => {
                          'actin' => '20,514',
                          'claudin' => '30,544',
                          'occludin' => '31,183'
                        },
          'H2O' => {
                     'actin' => undef,
                     'occludin' => undef,
                     'claudin' => undef
                   },
          'Sample 4' => {
                          'actin' => '20,498',
                          'occludin' => '27,869',
                          'claudin' => '25,365'
                        },
          'Sample 3' => {
                          'occludin' => '34,556',
                          'claudin' => '32,065',
                          'actin' => '24,213'
                        },
          'Sample 2' => {
                          'claudin' => '25,611',
                          'occludin' => '27,831',
                          'actin' => '20,416'
                        }
        };
[download]

I am wondering though if you want to right on on top of common keys or you want a more complex data structure like hashes of hashes, read more about it here HASHES OF HASHES.

Hope this helps, BR.