8 comments

[ 4.6 ms ] story [ 27.6 ms ] thread
Interesting stuff but light on details. Anyone work in this field who can provide insight as to how the data is initially encoded in the DNA?
Complicated to explain without being able to refernce the figures, but they basically had "symbols" that consisted of short DNA sequences. These were followed by longer sequences that made a "tail" of DNA.
And what the decoded representation is...
they are about to take over...the end is near :|
Please link to the actual research, not the journalist's article.

http://www.scripps.edu/news/press_releases/20120207keinan.ht...

That being said, am I missing something here? Information encryption and image encoding are two entirely separate things.

I'm not paying for the research paper, but the abstract states:

> Parallel computing with molecular finite-state automata and fluorescently labeled DNA molecules has been used to decipher two different images encrypted onto a single DNA chip (see picture). The images were deciphered by a mixture of input molecules that were processed by biomolecular automata, a strategy that potentially offers a huge diversity of encrypted images.

Again, this isn't encryption as I understand it in the traditional 3DES/AES/pgp sense. This is something closer to stenography or encoding/decoding.

Can anyone clear up what exactly is going on here for me?

He probably didn't link to a real article because he works for "extremetech.com" since all his submissions history are articles from there.
There's a guy in my deptartment that does stuff like this. The basic idea is to model chemical reactions with DNA. The chemical reactions surprisingly simple (x + y' --> y is essentially an AND gate, for example). In fact, they've even made a C-->chemical reaction compiler!

>the jury is still out on applications

I disagree -- I see huge applications, especially in medicine. There were some people from Caltech that used a DNA-computer to kill cancer cells extremely effectively.

More papers are available here: http://cctbio.ece.umn.edu/wiki/index.php/Papers,_Theses,_and...

The compiler: http://cctbio.ece.umn.edu/chem-compiler/chem-compiler.pl

Cancer application: http://www.wisdom.weizmann.ac.il/~udi/papers/automoleculcomp...

> We’re not talking about a molecular computer that’s comparable to the CPU in your PC, though; rather, the scientists created a simple Turing machine-like finite state automaton.

This part made me laugh a little. The reporter knows what a Turing machine is, but doesn't know about the Church-Turing thesis. Not surprising.