I felt that way last week reading about the emergence of an unexpected new merger of biology and engineering: DNA computers. Scientists have discovered a way to take pieces of DNA-the stuff of life, genetic material found in the cells of every living creature-and use them as,“molecular computers” that have the potential to dwarf the powers of the mightiest supercomputers. One expert guesses that a system using a jug’s worth of DNA goop can crunch more arithmetic than all the conventional computers on planet earth. On one hand this is rather mind-blowing. But it all makes sense when you consider what really makes the molecular computers powerful: they use the same techniques that nature has been hacking with for millions of years. It’s like Stallone and Schwarzenegger teaming up in the ultimate buddy picture-biology and electronic computers, together at last. The future may never be the same.
The field of DNA computation was kick-started by Leonard Adleman, a 48-year-old USC mathematician. He began with the explicit goal of investigating links between biology and computation: “the two scientific disciplines that have burgeoned in the last half century,” he explains. To date, our approaches in those fields have by necessity been quite divergent. Biology presents us with fully formed masterpieces-we tenderly attempt to deconstruct them, and have only begun to deeply understand their intricate latticework. In the case of computation, we build from the ground up. Only occasionally, as in the case of genetic algorithms, have computer scientists seriously raided nature’s bag of tricks. Adleman attempted something particularly ambitious: using synthetic DNA as the computer itself. When he got his molecular machine to solve a mildly challenging math puzzle, he considered it more a proof of concept than an industrial breakthrough. But since publishing his paper late last year, Adleman has been surprised at how quickly people are focusing on the practical implications of his discovery.
Why do molecular computers have so much potential? Because, with very little cost and almost no output in energy, they use the same microscopic massive parallelism that’s found in nature; every one of your cells, for instance, can be seen as a single processor, contributing to the ultrasmart supercomputer that is you. Biological calculations are slower than lightning-quick operations that occur in computer chips, but nature provides so many calculators that organic chemistry beats Cray routinely. Consider the problem of keeping the turf trimmed on a large golfcourse. The conventional solution involves a few power mowers. Each mower spends very little time on an individual blade of grass, but there is so much territory to cover that it takes a while to do the whole job. (Similarly, a powerful computer tackling a very complex mathematics problem handles more than a million calculations a second, but still can take days to solve it.) But what if you were presented with billions of tiny lawn mowers that took an hour to slice each strand of grass? Assign each mower to a single blade of grass, and the job is done in an hour. That’s how a DNA computer attacks a task. A real-world problem, like sorting through trillions of possible digital keys to find the single string of numbers that “unlocks” the text in a message encoded by the Data Encryption Standard (DES), no longer requires a computer affordable only by Bill Gates or the U.S. govemment. Using a rack full of test tubes, a glass full of DNA and the proper techniques, “you could almost do it in your basement,” says Princeton computer scientist Richard Lipton. Voila: you’ve cracked the code that banks use to protect your financial data.
Don’t close out your savings account quite yet. All of the scientists working in this nascent field offer caution in estimating how quickly it will extract such results -or whether the whole thing won’t prove a chimera. Even in the most optimistic scenario we won’t be hacking away at laptops bearing stickers that say DNA INSIDE-most of the conventional uses of computers can’t benefit from the massive parallelism that bestows magic upon molecular computing. Still, no matter how these experiments pan out, the original breakthrough leaves us with solid lessons.
The first is a continued respect for the true master of computation-nature. While synthetic DNA computing appears to be marvelous, it’s toyland compared with what goes on in our bodies every minute, where a handful of genetic material runs an amazing juggling show with billions of cells acting in seamless coordination for the good of the organism.
Second is the humbling realization that trying to nail down the future is a fool’s errand. While we’re struggling to assimilate the effects of the breakthroughs we’ve already seen, those damn innovators keep coming up with new ones. If biology and computers truly join forces, though, the changes will probably be so pervasive that no one will be able to escape them. The speculation ranges from supermedicines to nanotechnological “assemblers” that rearrange atoms as easily as Lego blocks. So brace yourselves -that big whooshing you hear may well be the sound of the nucleotides turning.
