You should be using CGI for this, but see if this helps you (untested):

print "<table>\n";
my %names;
{
  open my $name, 'gly_structures.txt' or die;
  while ( <$name> ) {
    my ( $ref_num, $glycan_name ) = split /\|/;
    die if defined $names{ $ref_num }; # die on duplicates
    $names{ $ref_num } = $glycan_name;
  }
}

open my $struct, 'N_structures.txt' or die;

while ( <$struct> ) {
  my $number = ( /(\d+)/ )[ 0 ];
  next unless defined $number;
  my $GU_web;
  open my $info_web, "${number}_info.txt" or die;
  while ( <$info_web> ) {
    if (/^(\d+\.\d+)\D/) {      # probably wrong
      $GU_web = $1;
      last;
    }
  }
  close $info_web;
  die unless defined $GU_web;

  my $image = qq(<IMG SRC="images/$number.gif">);
  defined( my $structure_name = $names{ $number } ) or die;

  print "<tr><td>$GU_web</td><td>$image</td><td>$structure_name</td></
+tr>\n";
}
print "</table>\n";
[download]