In the beginning, there was ENIAC. The first electrical computer could do 5,000 additions or subtractions per second, recounts Mark Reed, as long as people with shopping carts full of vacuum tubes jumped to the rescue each time the behemoth suffered a burnout. Then came transistors, and integrated circuits, greatly increasing the number of operations machines could perform, even as their components shrank. But now researchers face a serious barrier in miniaturization, called power dissipation.

As technologies scale down, and more computer chips get packed together, the number of watts per square centimeter reaches a point “when materials start to do nasty things, like break down,” says Reed. To break through the power dissipation barrier Reed and others look to biologically inspired systems. DNA could prove the ultimate scaffolding for new computational structures, believes Reed. He shows some examples of DNA folded like origami, and assembled into such patterns as stars, and even a map of the world. So why not create a simple component, like a switch? Researchers have fashioned RNA into just such a device, providing input signals via metal atoms, proteins, and other simple chemicals. They have even figured out how to send the signal from one artificial biological structure to another. This “is not far afield from the typical input/output of modern computers,” says Reed.

Scientists are developing a new generation of biosensors that can detect an electrical signal that might emanate from the smallest building blocks of life. Reed has developed “nanowires,” and applying etching techniques like those used for current semiconductors, fashioned biosensors that can detect minute changes in various biochemical environments. “Not only can we measure things like DNA and other proteins, but we can also talk to cells,” says Reed. The nanowire sensor has a future as a diagnostic tool, because it can read the biochemical messages cells send out when they’re sick -- in real time.

Eventually, Reed sees integrating complicated DNA structures with electronics that look like those found in our computers, and devising sensible interfaces to the biochemical world, “involving two-way communication with this environment.” Computation using such structures may start out slow, but massively parallel processing could bring speeds up without using a lot of power.