Perl can be viewed as an implementation of sed, tr and awk, along with many functions of the shell and of C. You won't need to run command lines from a perl script in order to do tr and sed work, because that sort of thing is an integral part of (and much more capably implemented in) Perl.

In fact, you won't want to run commands for that kind of stuff, especially on a "one command per word" basis, because the overhead of launching and closing down a new shell to run such a command for every word will slow you down tremendously.

Just use the "tr///" and "s///" operators in Perl, along with the standard IO operations to read/write lines, and split to divide each line into individual words.