Alien Genetic Code
Variable Length Genetic Codes

In English, words are defined by boundaries, for example, by spaces and certain punctuation. Words may be of any length, but they may not contain the symbols, like spaces, that delimit them. That's not the way our genetic code works, but one can imagine an alien genetic code that does work that way.

For example, suppose that 'space' in an alien genetic code is represented by the letter 'C'. Then an RNA sequence ...AGCCGUUCACUAGUCUCAUC... may be read as: 

     AGCCG UU A UAGU U AU
So long as the resulting codons UU, A, UAGU, etc are defined in this code, the message would make sense. No codon, however, could contain the space symbol C.

The delimiter may also be more than one letter. Suppose that it is UC, then the same message could be read as: 

     AGCCG U ACUAG A
Now, no codon is permitted to contain UC, but C's are OK. Note that translation in your in-vitro system (unlike natural translation) can initiate anywhere in the RNA (see IN-VITRO-TRANSLATION).

In the absence of experimental results, any of these possibilities (and many more) may be right. However, you can exclude many of them by assuming them to be true and showing that doing so leads to a contradiction, for example a codon that codes for one amino acid in one context and another in a different context.