They also may not publish the DNA sequence they are using, or they may have more than one sequence and only publish one of them. Presumably it has to be in a highly conserved region of the bacterial genome (or present in a few highly conserved places).
From what I understand they use insertion sequence element[1] combined with a given set of restriction enzyme to produce a fingerprint. They had to find strains that produced a unique fingerprint given their set of restriction enzyme.
> they may have more than one sequence and only publish one of them
In fact that's exactly what they do. From the "7. Risk of Duplication of the Proof-of-origin Culture" section of the paper:
> And last but not least, the proof-of-origin culture consists of more than one traceable strain.The composition of the mixture is changed (rotation principle) from time to time in accordance with the interprofessional cheese organization.
