This article is more than 1 year old
Boffins put encrypted bio-copyright watermarks in beer DNA
No knocking off Blade Runner pleasure replicants
German boffins believe they have developed a computer algorithm which can be used to hide encrypted "watermarks" within the DNA of living genetically-modified organisms. The procedure has been successful in simulated tests on live beer ingredients.
Dominik Heider and Angelika Barnekow of the Department of Experimental Tumorbiology at the University of Muenster explain their techniques in an academic paper released last month.
The two boffins' research is aimed at "the application of watermarks based on DNA sequences to identify the unauthorised use of genetically modified organisms (GMOs) protected by patents". The idea is that a patented GM organism - such as a crop, a drug or perhaps in future a Blade Runner-style engineered human replicant - might be pirated by unscrupulous rival manufacturers, who could then produce ripoff copies without doing any development work.
Not if Heider and Barnekow have anything to do with it, though. Dodgy supermarket-carpark pleasure clones or whatnot would be easily identified as branded product using the DNA watermark, hidden among the information in their cells just as a microdot holding a hidden page of text can masquerade as a full stop on a sheet of paper.
Of course, without encryption, the ripped bio-products could be easily given fake branding in the same fashion as a knockoff Rolex. The German boffins' bio-stegano-cryptograms, however, take that into account. Rather than an obvious image or text, the hidden DNA info would be encrypted. Their computer program, DNA-Crypt, can be combined with binary encryption algorithms like AES, RSA or Blowfish, or can be used with one-time pads.
Apparently that's fairly yawn-worthy in the world of biosteganocrypto-boffinry - Heider and Barnekow cite several previous researchers who've hidden encrypted messages in DNA. The Germans' special sauce is that their DNA-Crypt program can deal with the occurrence of mutations, in which the DNA of the organism in question changes unpredictably as it reproduces.
"Mutations do not occur very often, approximately 10−10 to 10−15 per cell division, but they can destroy the encrypted information in DNA sequences," according to the Muenster scientists. If a cop or future Replicant-Industry-Association-of-America (RIAA) enforcer checked a mutated sample, the watermark could be reduced to hash and the bio-ripper might get off scot-free.
But the biocopyright-loving boffins reckon they've dealt with this, using "the 8/4 Hamming-code and ... the WDH-code," which are methods of writing to DNA which can provide "not only ... error detection but error corrections which enable us to maintain the data." These methods use up more space than ordinary DNA fiddling, so the DNA-Crypt platform uses an "integrated fuzzy controller" which "decides and recommends whether to use the 8/4 Hamming-code, the WDH-code or no mutation correction for optimal performance." We were especially pleased to hear that it "uses the Singleton-fuzzyfication," which ought to be a great marketing tool if nothing else. ("Nexus 6 pleasure models, verified genuine by DNA-Crypt™:now with Singleton fuzzyfication for optimal performance.")
Heider and Barnekow have done successful tests of their procedure on Saccharomyces cerevisiae, better known (and loved for its beautiful effects) as brewer's yeast. Less importantly, it's also used to make bread. However, the watermarked brewing yeast trials were only in silico - in computer simulations. No real-world biowatermarked yeast, let alone beer or pleasure/warrior replicants, has yet been produced.
For those interested, the DNA-Crypt code is Java-based (5.0 and higher) and cross-platform: Mac or Linux-using replicant designers can get in on the biowatermarking action.
A preliminary pdf of the research paper is here, and the project homepage - which will apparently offer DNA-Crypt for download in future - is here.®