Friday, November 21, 2003

DNA

We just got one step closer to the age of nanotech. As reported in New Scientist, researchers have used E. Coli bacteria DNA and carbon nanotubes to self-assemble transistors. This could be the limit for the low end of transistor size. There is a significant amount of work that must be done before they can self-assemble entire circuits, however... combining many of these transistors together to make a useful device will likely involve protein folding techniques, which are some of the most intractable problems in science at this time.

Researchers have already used DNA to self-assemble patterned nanostructures, circuit elements, and a sort of nanodevice scaffolding that could be the first step in the creation of nanotech machinery. These approach the size limit for any physical device... building devices one molecule at a time, in contrast to almost all other small-scale assembly techniques which remove matter until the desired structure is all that remains.

The inherent ability of DNA to assemble complicated structures (like you, for instance) has also been used to solve mathematical problems such as the Hamiltonian path or 'travelling-salesman' problem and prime number factorization, the latter of which is of critical importance in cryptology... almost all 'secure' computing relies on the near-impossibility of finding the prime factors of absurdly large numbers in a short enough time that the encrypted data is still relevant. (At today's computing speeds, some of the algorithms would take tens of thousands of years, or a really lucky guess, to crack).

DNA isn't particularly fast. DNA computers (as seen on the level of logic gates) are therefore not very fast. But the advantage of DNA computing is that billions of very complex operations can take place simultaneously in a test tube, and well-understood methods can then be used to extract the correct answer. Our brains are slow: some neural process take on the order of milliseconds, whereas the computer I am now typing on can perform a single operation in half of a nanosecond. But the parallel structure of our brains makes short work of many complex problems. Nature has evolved an incredible supercomputer (hell, nature is an incredible supercomputer); I'm glad to see that we are starting to incorporate some of that brilliant evolution into our own creations.