This is interesting.
Carbon nanotubes have properties that make them very interesting as possible channel materials for advanced transistors. Unfortunately, precise placement of trillions of nanotubes is a difficult problem.
Enter DNA. DNA is very good at self-assembly, so it's (relatively) easy to make an array of DNA structures. It's also easy to modify a DNA molecule so that it will bind to, say, a carbon nanotube.
At least that's the idea behind recent work at IBM. So far they're still working on the DNA scaffold, but the potential is very cool.