Hello Anonymous Monk

Maybe you are right let's Benchmark them and see:

I know that the while loop exceeds the foreach loop but in my case I did not made many differences but I was aiming to remove unnecessary lines.

#!usr/bin/perl
use strict;
use warnings;
use Benchmark::Forking qw( timethese cmpthese ); # UnixOS
# use Benchmark qw(:all) ; # WindowsOS

my @starts;

sub previous {
    open (FASTA, "sequence.fa")
    || die "Cannot open file: $!.\n";

    chomp (my @seq = <FASTA>);

    close FASTA;

    shift @seq;

    my $sequence = join ('', @seq);

    @seq = split ('', $sequence);

    for (my $i=0; $i<=$#seq-5; $i++){  ## -5 könnte man weglassen
    # start codon: ATG
    # stopp codon: TAA, TGA, TAG
    # multiple of 3 between start and stop
    if ($seq[$i] eq 'A' && $seq[$i+1] eq 'T' && $seq[$i+2] eq 'G') {
        push (@starts, $i);
        for (my $j=$i+3; $j<=$#seq-2; $j=$j+3){
        if ( ($seq[$j] eq 'T' && $seq[$j+1] eq 'A' && $seq[$j+2] eq 'A
+') ||
             ($seq[$j] eq 'T' && $seq[$j+1] eq 'G' && $seq[$j+2] eq 'A
+') ||
             ($seq[$j] eq 'T' && $seq[$j+1] eq 'A' && $seq[$j+2] eq 'G
+') ) {
            # print "ORF: $i-", ($j+2), "\n";
            last;  ##lasts the j loop
        }
        }
    }
    }
    return;
}

sub update {
    open my $fh, "sequence.fa"
    or die "Could not open file: $!";

    while (defined( $_ = <$fh>)) {
    chomp;
    next if $. < 2; # Skip first line
    my @seq = split '', $_;
    for (0..$#seq-5){
        if ($seq[$_] eq 'A' && $seq[$_+1] eq 'T' && $seq[$_+2] eq 'G')
+ {
        push (@starts, $_);
        for (my $j=$_+3; $j<=$#seq-2; $j=$j+3){
            if ( ($seq[$j] eq 'T' && $seq[$j+1] eq 'A' && $seq[$j+2] e
+q 'A') ||
             ($seq[$j] eq 'T' && $seq[$j+1] eq 'G' && $seq[$j+2] eq 'A
+') ||
             ($seq[$j] eq 'T' && $seq[$j+1] eq 'A' && $seq[$j+2] eq 'G
+') ) {
            # print "ORF: $_-", ($j+2), "\n";
            last;  ##lasts the j loop
            }
        }
        }
    }
    } continue {
    # close ARGV if eof; # reset $.
    }
    close $fh
    or die "Could not close file: $!";
    return;
}

my $results = timethese(1000000, { Previous => \&previous,
                   Updated => \&update }, 'none');
cmpthese( $results );

__END__

$ perl bio_test.pl
           Rate Previous  Updated
Previous 2224/s       --     -14%
Updated  2601/s      17%       --
[download]

See also my update proposed solution, the second update should resolve the question and should be also faster.

The O(n**2) nested-loop performance is going to kill you on some datasets. For a really pathological one, try:

$sequence = 'ATG' x 1e6;
[download]

I estimate that your code would take about a day and a half to process that. My code handles it in just over a second.

The human genome is around 3e9 base-pairs long. That's small enough to fit it all in memory, but large enough that you need to use efficient algorithms on it.