Hello,

Since you already know that your fasta string can only contain characters A,T,G,C other than N, the simplest way is to just use that bit of information. :-)

Your regex is fine, but the problem is that it can match 'N' anywhere in the string. Here's how we can use our little tidbit to advantage. Change your regex to: [ATGC]N+[ATGC].

Here's some sample code demonstrating it:

use strict;
use warnings;

while (my $line = <DATA>) {
chomp $line;
# The below regex tells perl to look for
# any of A,T,G,C followed by a string of 
# one or more Ns, followed by A,T,G,C.
my @info = split /[ATGC]N+[ATGC]/, $line;
print join(", ", @info), "\n";
}

__DATA__
NNNAAAATATGACAAAGGGGTTNNNNNNNNNNNNNNGATGTCTGGTCAATAGGAT
CGCAGCCATTAACATCTCAACAAGCCAAAAATTCCTTCTCAGAAATTCGGNNN
AAAATATGACAAAGGGGTTNNNNNNNNNNNNNNGATGTCTGGTCAATAGGAT
[download]

Update: My split solution has a problem in that it loses one of [ATGC] on either side of the internal pattern. Please use this solution by hdb or johngg's extractive matching. If you prefer to use lookaround assertions, here's one by oiskuu.