Design By DNA

Check out Spencer Reiss' cool interview with NYU chemist Ned Seeman from the lastest issue of Technology Review magazine. Seeman believes, as I do, that DNA molecules could be a perfectc assembly platform for the smallest and most powerful computing devices ever built.

How do you build things out of DNA?
We don't. DNA is just a way of organizing materials on a molecular level. It's scaffolding. For instance, carbon nanotubes--how are you going to organize them into a circuit? DNA gives you a way to arrange them into something useful. Because it has a very precise structure, and because you can control how other molecules associate with it, it's just punching a sequence into a machine. And because DNA self-assembles, if there are things attached to it--micro metallic particles or carbon nanotubes--those will self-assemble along with it.

DNA's a linear molecule. Why doesn't everything you make wind up being linear?
We use a synthetic form, which we program to give us branch points. Think of the double helix as two lanes of a highway; branched DNA corresponds to intersections. You can make molecules of pretty much any shape or size you want.

What kinds of things have you made?
Lots of crystals. The earliest complex device was something that changed its shape in a controlled fashion when you added a chemical. Last summer, we did a little walker that moved across a DNA "sidewalk." Each foot was tied down by a strand of DNA. We would rip off that strand, and then the foot was free to wander around, and then we'd put in another strand to tie it down and make the next step.

How does computing come in?
As things in the computer world keep getting smaller, they're reaching the point where top-down approaches--trying to make big things smaller--are hitting the wall. What we're doing is building from the bottom up--taking little things and make them bigger. And DNA lets you do true 3-D integration. There are issues of cooling and power loss that have to be addressed, but the point is that what we're doing is inherently three dimensional, which at the nano level is pretty amazing.

So is nanomanufacturing imminent?
We are probably not going to be using this approach to knit customized sweaters. DNA is expensive stuff; for now, at least, you wouldn't want to use it for large-scale anything. But
3-D configurations of atoms, or molecules, or nanoparticles--that has to have value, in terms of making things no one has been able to make before.

What about nanotech's skeptics?
Everything we're talking about is doable. Is it doable on a scale that's going to be worthwhile? No one knows. In 25 years we've taken something that was in my imagination to the point where we can take out patents and where there are now whole conferences devoted to the topic.