The expense of decoding DNA has now plummeted

Fifteen years ago, gels and fluoreseenl dyes lay at the heart of every machine that sequenced DNA

MARCO ISLAND, FLORIDA—Fifteen years ago, gels and fluoreseenl dyes lay at the heart of every machine that sequenced DNA. Bases cost a dollar or more to sequence, and deciphering a human genome would take years. The expense of decoding DNA has now plummeted, and new human genomes appear in quick succession thanks to advances in DNA-sequcncing technologies and a growing roster of sequencer manufacturers. Whereas one company dominated the industry 6 years ago, about a half-dozen companies now produce DNA-sequencing machines and novel technologies come onto the scene almost annually. Attesting to this revolution was a packed house at the final Saturday session of a genome meeting here, where silicon wafers and quantum dots were the DNA-sequencing technologies du jour.

Although neither approach is yet able to fulfill the dream of the $1000 human genome, these newcomers have the potential to change the face of DNA sequencing and expand the ability of researchers, and eventually clinicians, to incorporate DNA into their work. "It's pretty exciting that the field has opened up as much as it has," says Elaine Mardis of the Genome Center at Washington University in St. Louis, Missouri. At the same time, labs pursuing a sequencing project face ever more choices and tradeoffs between speed, accuracy, and cost. With so many technologies to choose from, "it's confusing," says Eric Green, director of the National Human Genome Research Institute (NHGRI) in Bethesda, Maryland. This year's upstart at the annual Advances in Genome Biology and Technology meeting here was Ion Torrent Systems Inc., based in Guilford, Connecticut. Started 2 years ago by Jonathan Rothberg, who founded and sold the sequencing company 454 to Roche in 2007, Ion Torrent offers a DNA-sequencing strategy that's "a radically different proposal," says Edward Rubin, director of the Department of Energy Joint Genome Institute (JGI) in Walnut Creek, California.

In current "next generation" DNA sequencers, such as those made by 454, the machine decodes a strand of DNA by using it as a template to synthesize a matching strand, where base additions are signaled by the emissions of photons of light. A camera records each base, revealing the corresponding one on the template DNA. In most cases, the original DNA must be cut into small pieces that are copied many times over before being sequenced. Those pieces are anchored on beads or slides, and sequencing is done in a massively parallel fashion, achieving rates unthinkable with the gel-based technology used to decode the first human genome.

In contrast. Ion Torrent's sequencer is a silicon chip, built the same way as a semiconductor and etched with an array of nanoscopic wells 400 per the width of a human hair. The wells sit on top of an ion-sensitive layer a pH meter of sorts- below which is the layer that transmits electrical current. The DNA to be deciphered goes into the wells. Polymerase, the enzyme that builds up a matching strand, is added, along with each of the four different bases, one type of base at a time. When the polymerase finds the right base in the well and attaches it to the new DNA, the reaction releases a hydrogen ion that the chip detects as a voltage change. If the base in the well is wrong and isn't added, no voltage change happens. Then unattached bases are washed out and another type of base is added. Ultimately, the series of electrical pulses recorded by the chip translates into a readout of the DNA being sequenced.

To date, Ion Torrent has sequenced the genomes of only a virus and a bacterium. But according to Rothberg, each took about an hour, 100 times faster than some of the next-generation machines now on the market. At this point, the chip-based system cannot do high-volume sequencing, says NHGRI's Jeffery Sehloss, "but it has the potential of filling a very important niche" for small-scale, quick-turnaround jobs.

Rothberg expects to greatly increase the density of the wells and thus the efficiency of sequencing. According to Ion Torrent, its machines will be for sale by the end of the year, priced to put them in reach of many smaller biology labs. "It's one of the besi marketed and coolest concepts of where sequencing could potentially go," says Len Pcnnacchio of JGI.

But it wasn't the only concept to create a buzz at Marco Island. Life Technologies in Carlsbad, California, has in the wings a new strategy aimed at sequencing a single DNA molecule most other technologies must analyze multiple copies of the DNA to be deciphered. Life Technologies' approach involves tethering a nanocrystal semiconductor called a quantum dot to a DNA polymerase. A laser excites the dot, which then transfers energy to fluorescent dye-tagged bases, but only when the polymerase adds a base to the DNA chain being built from the template being sequenced. A camera detects each added base by the color emitted from its dye, immediately translating that data into the sequence of the template.

At this point, says Joseph Beechem of Life Technologies, the company has tried the technology only on humanmade test DNA strands. And the approach is not perfect, as the quantum-dot nanocrystals blink on and off and might miss a base addition, says Mardis. Still, she and others agree that the technology has the potential to decipher long stretches of DNA in a single sweep. Life Technologies' machine might eventually challenge a single-molecule sequencer produced by Pacific Biosciences, based in Menlo Park, California. But the latter has a head start: At the meeting, it unveiled its first commercial machine with much fanfare, including fireworks on the beach.