Here's a simpler Perl solution (similar to the Awk solution of Frédéric Mahé):

$ echo -e 'gene1 contig2 1 69 100 169\ngene2 contig20 3 53 250 200' \
| perl -ane 'print join "\t", @F, $F[4] > $F[5] ? "-\n" : "+\n"'
gene1    contig2     1    69    100    169    +
gene2    contig20    3    53    250    200    -

You could make it perhaps more readable by adding explicit loops and variables, but for one-liners I think it's best to use the tools you have and save yourself some typing.

EDIT: Perl's command line switches are documented in perlrun (typeperldoc perlrun from the command line).

• The -e tells Perl to process the command line arguments, which would be any files or STDIN (as is the case above).
• The -n switch will make Perl loop over the input line by line (the -p does the same, but turns on an implicit print).
• The -a tells Perl to autosplit the input and put it into an array called "@F" when used with -n or -p. You can change the delimiter with the -F switch.

Simple awk solution awk '{if ($5>$6) print $0,"-"; else print $0,"+"}' INPUT

echo -e 'gene1 contig2 1 69 100 169\ngene2 contig20 3 53 250 200' |  awk '{if ($5>$6) print $0,"-"; else print $0,"+"}'
>gene1 contig2 1 69 100 169 +  
gene2 contig20 3 53 250 200 -