Meller has not yet demonstrated his sequencing technology, but he says it relies on a combination of proven techniques and, unlike electrical nanopore sequencing techniques, there are no fundamental problems remaining. His group has fabricated 25-nanopore arrays and demonstrated the effectiveness of the fluorescent probes. Now, with a three-year, $2.2-million NIH grant, he believes they'll be able to put the pieces together.

If this technique works, it could mean DNA sequencing at extremely high throughput. Meller says arrays of 100 by 100 nanopores could sequence an entire human genome in an hour.

Current sequencing techniques rely on many repeated biochemical steps, including replicating the DNA, adding probes, and chopping it into pieces. It takes weeks for computers to process the resulting sequence fragments and piece them back together. Moreover, the computer processing is expensive and cannot piece together long sections of the genome in which the same sequence is repeated. Today's techniques also cannot be used to find rearrangements--movements of large stretches of DNA from one part of a chromosome to another. Yet rearranged and repeated segments are believed to be important elements of genomics.

"Research is beginning to show that long-range rearrangements have disease effects," says Jeffery Schloss, director of technology development at the genome institute. Yet current sequencing techniques cannot get at these. "If [nanopore sequencing] works at all, you ought to be able to sequence long stretches" because the DNA is read continuously in real time, he says.

Another exciting possibility, according to Schloss, is using nanopore sequencing to read molecular modifications to DNA that regulate gene expression. Inside cells, for example, genes are sometimes covered in hydrocarbon molecules, called methyl groups, effectively silencing them. (It's how females operate with two X chromosomes: in each cell, one is covered with methyl groups and inactivated.) While traditional sequencing strips off methyl groups and other modifications, this step may not be necessary using nanopores. "I really like the idea of being able to read the native DNA," says Schloss.