use strict;
use warnings;

my $str='ABC---DR----EEEEEGGGG-GRE-RED----KKKK---';
my $len_seq = () = $str =~ /[^-]/g;
[download]

This is a FAQ:

perldoc -q count
[download]

See also:

• How can I count the number of occurrences of a substring within a string?
• Perl Idioms Explained - my $count = () = /.../g

In addition, you could do:

my $len_seq = $str =~ tr/A-Za-z//;

Or possibly:

my $len_seq = $str =~ tr/-//c;

[download]

Another way, also without modification of the original string, perhaps slightly faster:

>perl -wMstrict -le
"my $str = 'ABC---DR----EEEEEGGGG-GRE-RED----KKKK---';
 print qq{'$str'};
 my $len = $str =~ tr/A-Za-z//;
 print qq{'$str'};
 print $len;
"
'ABC---DR----EEEEEGGGG-GRE-RED----KKKK---'
'ABC---DR----EEEEEGGGG-GRE-RED----KKKK---'
24
[download]

You can save a little typing and also be a little more comprehensive with respect to "anything not a dash" by using the c flag (complement) of tr, see Quote and Quote like Operators.

knoppix@Microknoppix:~$ perl -E '
> $str = q{ABC---DR----EEEEEGGGG-GRE-RED----KKKK---};
> $len_seq = $str =~ tr{-}{}c;
> say $len_seq;'
24
knoppix@Microknoppix:~$ 
[download]

I hope this is of interest.

Cheers,

JohnGG