IN THE DECADE since researchers first sequenced the human genome, obvious links between the genes and individual diseases have been slow to appear [see “Revolution Postponed,” by Stephen S. Hall; Scientific American, October]. Many researchers now believe that real advances in genomics will come not from simple X-causes-Y correlations but from a rich statistical understanding that emerges out of the sequences of millions of genomes—a set that reveals how our genetic code is likely to interact with the environment to make us who we are.

This, in turn, requires a cheap genome sequencer, something that can do the job for less than $1,000. Right now it costs between $5,000 and $15,000 to sequence a genome—a great improvement from the $2.7 billion it originally cost but still far from the goal. Researchers at IBM and Roche are trying to get there by undertaking a radical redesign of gene-sequencing machines. Whereas existing dishwasher-size sequencers require expensive chemical reagents to analyze genes that have been sliced into thousands of small fragments, the so-called DNA transistor takes an almost naively simple approach. In it, an intact DNA molecule threads through a three-nanometer-wide gap in the middle of a silicon chip. As the DNA feeds through the nanopore, an electrical sensor reads it one molecular unit, or base, at a time.

Other labs have experimented with similar nanopore-based approaches to sequencing but have found it difficult to control how quickly the DNA strand feeds through the nanoscale hole. The IBM team has hit on a method that capitalizes on DNA's naturally occurring negative charge. “We thought that if the device contained electrodes in the pore itself—thin layers of metals separated by insulating material—that electric field would interact with the DNA,” says Gustavo A. Stolovitzky, a research scientist working on the project. The electric field grabs the negatively charged DNA and holds it in place. When the electric field shuts off, the strand continues to move through the hole until the next base lines up for sequencing. At that point the electric field reappears, and the process repeats itself all the way down the strand.

The technique isn't a slam dunk. The pore must produce a strong electric field to hold the DNA in place. But the high voltage needed to create this field can cause what is known as a dielectric breakdown, where sparks fly and the electric field shorts out, which is especially likely to happen over such short distances. “It's as if you have a cloud very close to the earth—it's much easier for lightning to strike,” Stolovitzky says. The researchers are looking for an electrode material that can withstand the necessary charge.

Despite these issues, industry observers think the DNA transistor can be a fast, cheap, efficient way to sequence genomes. “This reduces the number of steps required for sequencing—it's just literally looking at the DNA itself,” says Bruce Schiamberg, a consultant who evaluates the commercial potential of biotechnology innovations. “There are no reagent costs or optical instruments needed to read fluorescent tags. The thing gets done faster.”

The DNA transistor is on track to supply a complete genome sequence within the next few years for less than $1,000. Stolovitzky believes the device will help the scientific community more readily make connections between genes, health vulnerabilities and ideal drug treatments. With a statistical understanding of the connections between genes and disease, pharmaceutical companies could better target drug development, because they would already know what regions their new drugs would need to focus on. He points to one of the early success stories: herceptin, a breast cancer treatment that halts tumor growth in patients who show overexpression of a gene called HER2. “There are a handful of these examples,” Stolovitzky says. “We would like for it to become a very common thing.”