top | item 2613489

DNA Logic Gates Calculate a Square Root

63 points| jedi_stannis | 15 years ago |arstechnica.com | reply

14 comments

order
[+] ChuckMcM|15 years ago|reply
To my way of thinking I'd like to see a species of grass that factors 2048 bit integers. Figuring that once a factorization was found it would express a gene which turned to color of the blade red, then every week after you mow you lawn you could look through your clippings to find the red ones, sequence those and then see if you had the keys to any banks. :-)
[+] wgrover|15 years ago|reply
Nice! If you'd like to write biopunk sci-fi books, I'd read them. :)
[+] rcthompson|15 years ago|reply
So you'd be running a blade server farm?
[+] SpikeGronim|15 years ago|reply
You'd need a LOT of grass.
[+] catechu|15 years ago|reply
Brings new meaning to the field of computer science.
[+] fleitz|15 years ago|reply
Looks like it's time to buy farmland in Iowa.
[+] pjscott|15 years ago|reply
This is DNA strand displacement logic, a particularly simple (yet surprisingly robust) style of DNA computing. They can chain together "logic gates" made from reactions of partially complementary DNA oligonucleotides, computing arbitrary boolean logic functions, in an asynchronous logic style. The big potential here is that they can use this for drugs (and other biological purposes) that can do a bit of simple computation. Biology does this sort of thing all the time -- that's where it gets a lot of its power -- but this kind of simplified logic is easier for humans.

Looking over the paper, it looks like they figured out how to combine their seesaw gates and thresholding gates into AND and OR gates, then from there, used dual-rail asynchronous logic to make a four-bit square root calculator. They can't make NOT gates (or NAND, or NOR), so each bit gets two "wires", holding opposite values; this lets them make arbitrary logic without NOT operations.

There's a pretty slick simulator program you can play with here, along with more information, if anybody's interested:

http://research.microsoft.com/en-us/projects/dna/

[+] rcthompson|15 years ago|reply
It's neat, but are there any practical applications for this?
[+] ricree|15 years ago|reply
Various biomolecules, including DNA, RNA, enzymes, and small molecules, could all potentially be used as inputs. And it should be possible to link the outputs into relevant biological functions, including gene expression.

This would suggest, perhaps smart medicines capable of changing their active makeup based on logic in response to the actual conditions in the body.

Heck, given a lot of advancement it isn't that terribly outlandish to imagine simple turing complete computers capable of outputting different dna or even proteins. If that were possible, the benefits should be obvious. Of course, that's a rather large extrapolation from the limited results in the article.

[+] sliverstorm|15 years ago|reply
Obviously not at the moment, but the reason research like this happens is because as you develop the process it gets easier and faster, and who knows, in a few years it might get close enough to other processes to suddenly open new doorways you haven't even thought of.

As I understand it, plenty of research happens with no clear goal in mind, and plenty of research winds up good for something other than the original goal. See the Post-it note. This is, in my opinion, because you can't fully understand the technology until it actually exists!

[+] gcb|15 years ago|reply
Open publication. nice. take that old, dying, journals.