In Perl, how to speed up regex to modify a really big string?

Question

I have a string which comprised of 300 million bases;

$str = "ATCGTAGCTAGXCTAGCTAGCTGATXXXXATCGTAGCTAGCTGXTGCTAGCXXXXA...A";

I want to replace characters that are not [ATGC] in the string to something else, let's say to "A", meanwhile get the positions of characters that have be replaced;

I tried this:

while ($str=~/[^ATGC]/ig)
{
  $pos = pos($str);
  substr($str, $pos-1,1) = "A";
}

but the speed is not good.

Does anyone know better ways to do this?

Answer 1

Regexes can also substitute as well as match.

$str =~ s/X/A/g;

If you're only doing a single character, you can even use the tr operator.

$str =~ tr/X/A/g;

which may even be faster.

Answer 2

You can perform the replacement with regex directly using search and replace:

$str =~ s/X/A/ig;