MAGI should only get protein sequences as input. It is not designed to transform genes or RNA to protein sequences. (It can now translate codons to amino acids, but it does not recognise untranslated regions in proteins, such as ribosome binding sites or poly-A-tails or introns). It may be better to only accept protein sequences as input. This may also be clearer for a user.