Interesting idea here: as we're getting better at designing proteins for specific tasks, there might be a need to know who created what protein. This paper suggests adding 'watermarks'; in the context of proteins, this would seemingly involve changing the residue distribution according to a watermark code E derived from a private key.