Going from textual input to the psuedo base-4 system that DNA represents (CATG) doesn't seem optimal or effective at all prima facie, but there is more to it than meets the eye.
But nature, given millions of years of evolution, has fine-tuned its owned compression algorithm. Yet it only works on DNA - which is limited to 4 bases: C(ytosine), A(denine), T(hymine), G(guanine). The DNA replication process is error prone but there are heuristics we can use (getting multiple strands of our sequence being one) to detect errors. After you've generated DNA from an input, getting the data back out is significatly inefficient at this point, so the ideal use case of DNA would be archiving data that you don't need frequent access to.
There are three distinct steps:
