DNA sequencing technology has seen breathtaking improvements since we selected next-generation sequencing as Method of the Year in 2007. In 2011 we pointed to improved technologies that increased massive parallelization so that an entire human genome could be sequenced in a single run many times over, to a platform that no longer relied on optics for base reading but instead used ion sensing, and to improvements in single-molecule sequencing to yield very long reads. But even in a field as saturated with innovation as high-throughput sequencing, there is room for a yet more disruptive technology.

Sequencing through a pore. Credit: Marina Corral

The announcement of the first commercial nanopore sequencers by Oxford Nanopore Technologies at 2012's Advances in Genome Biology and Technology meeting raised great expectation in the scientific community—and for good reasons. Nanopore sequencing promises to deal with most shortcomings of current sequencing platforms: reads are very long, with tens, if not hundreds, of kilobases; errors, currently between 1% and 4%, are random rather than bunched together at the end of a read; data can be read in real time; throughput is high (the GridION, the bigger of Oxford Nanopore's machines, promises the capacity to sequence the human genome at 30× depth in less than a day); input DNA is not destroyed in the process; and sample preparation is simple and cheap.

In the sequencer model that is soon to be commercialized, a single DNA strand winds its way through a protein pore, its transition speed controlled by a second protein attached to the pore. As the bases traverse the pore, they alter the current that runs through the pore in a way characteristic for particular base combinations. Ideally one would see a typical pattern for every base, but the present machines will produce patterns characteristic for base triplets that then need to be computationally deconvoluted. And these patterns are not limited to the four DNA bases: nanopores can also detect methylated and hydroxymethylated bases and can, in principle, sequence RNA directly.

No doubt the community will carefully scrutinize nanopore sequencing data as they are released from the first testing sites to see how well these expectations are met.