#!/usr/bin/perl
use warnings;
use strict;

# firstly, put all data in hashes because it's easier to look for data
+ in a hash
my @input = ({},{}); # 2 hash references
my @output;
while (<>) { # this means: <STDIN> and/or filenames from @ARGV
        if (/^>([^\s]+)(\s+)?([^\s]+)?$/) {
                ++$input[0]->{$1}; # first match is an element of 1st 
+column
                defined $3 && ++$input[1]->{$3}; # third match might n
+ot exist and belongs to 2nd column
        }
}

# secondly, loop over first hash searching for elements in second hash
for (keys %{$input[0]}) {
        unless (exists ${$input[1]}{$_}) {
                push @output,$_;
        }
}

print join "\n",@output;
[download]