This article has been updated to clarify comments from Oxford Nanopore.
It's a fast-paced world. Sequencing technologies that are barely five years old are nearly passé and field-changing technologies from edgy startups are now the staples of larger companies. Next-generation sequencing technologies have opened up the field of genomics, allowing new avenues of research to be pursued than could be with Sanger sequencing. But these approaches are now under siege as the bleeding edge of sequencing technologies moves beyond sequencing-by-synthesis and toward single-molecule detection, microfluidics, and more — enabling researchers to do more sequencing reactions even more quickly.
The established sequencing companies — Illumina, Life Technologies, and Roche — are still honing their existing products while casting around for new methods and technologies to stay current. Right in their wake are the almost-to-the-market "third generation" sequencing approaches, from Pacific Biosciences, Complete Genomics, and others that also promise to shake up the field. But even those methods have competition on the horizon with new entries into the field based on microfluidics and nanopores. Everybody is in the race to the $1,000 genome.
Just which technology will be the first to reach that point is still anyone's guess. "It's really an open race at this point. It's not yet clear if a particular approach will win," says Cees Dekker, a professor at Delft University of Technology.
Currently, the sequencing marketplace is dominated by a handful of companies; while each has recently launched a new product, they are also looking toward partnerships, investments, and acquisitions to keep their edge in the sequencing market.
This year saw an influx of updated or scaled-down sequencers into the market. Life Technologies' SOLiD 4 machine came out in January with a list price of $495,000. According to the company, the machine can generate 100 gigabases of mappable sequence data per run and at an estimated reagent cost of $6,000 for a human genome. The SOLiD 4hq version brings those numbers down: 300 gigabases of sequence per run and an estimated $3,000 in reagent costs. "We've had good uptake of SOLiD 4 in the market," says Jay Therrien, the vice president of commercial operations for next-generation sequencing at Life Tech.
At the same time, Life Tech, Illumina, and Roche all introduced smaller versions of their sequencers to the market, mainly as a way to woo smaller labs.
Life Tech's $300,000 SOLiD PI has a slightly lower output than the SOLiD 4hq, but, Therrien adds, it has faster runtime and a more flexible flow-chip configuration.
Illumina's HiSeq came out earlier this year, priced at $690,000. It uses the same sequencing-by-synthesis chemistry as the company's Genome Analyzer, and can produce up to 200 gigabases of sequence data in about a week. It also uses the same cluster amplification sample prep that the GA does. "It's been terrific. It's actually quite exceeded our expectations," says Jay Flatley, the president and CEO of Illumina of the HiSeq's reception. "We've had more orders here in the first couple of quarters than expected and [are] a little surprised by [the] number of customers who have been able to get the additional funds to buy a HiSeq."
Around the same time, Roche came out with its GS Junior. It too uses the same method as its larger sibling, the GS FLX, and has a throughput of 35 megabases and a runtime of 10 hours. When the system was announced, it was said to cost about a fourth to a fifth of the price of the GS FLX, which goes for about $500,000. "Since its launch earlier this year, we have seen rapid uptake of the GS Junior System in the market," says Ulrich Schwoerer, who heads global marketing at 454. "The platform has been particularly well received in small- to medium-sized labs where access to funding for large enterprise sequencing platforms is limited."
In addition to new products, companies are turning to partnerships. Illumina has invested $18 million in Oxford Nanopore, bolstering its stake in the next wave of technologies. Under a 2009 commercialization agreement, Illumina will exclusively market, sell, distribute, and service Oxford Nanopore's exonuclease sequencing products. "We're very excited about nanopore sequencing as a concept. I think that is the best way to sequence, that we know of, conceptually," Flatley says. "We're very optimistic about how this technology could evolve over the next few years. It's not something that's ready for primetime today, but we believe that, in the long run, it is going to be very important in the next generation of sequencing beyond current systems."
Roche, too, is interested in nanopores.The company signed an agreement with IBM this summer to develop a nanopore-based sequencer based on IBM's DNA Transistor technology. Roche is funding the technology development, and has the exclusive right to market products based on the technology.
"We were particularly attracted to the IBM DNA Transistor technology because it addresses the fundamental challenge of nanopore sequencing: controlling the translocation of the DNA molecule through the membrane to enable precise detection of nucleotides," Schwoerer says. "Additionally, we believe that this approach, which uses a solid-state silicon nanopore, is more easily scalable than approaches which rely on biological nanopores."
Life Technologies has taken the acquisition route, ponying up $375 million in cash and stock for startup Ion Torrent, though that number may rise by $350 million if certain milestones are met.
A few companies and technologies are poised to burst onto the sequencing scene. This next wave is coming quickly and the technologies include refinements of sequencing-by-synthesis and single molecule sequencing, though the field is waiting impatiently to see whether these technologies — and the companies developing them — can deliver on their promises to speed up the rate of sequencing while bringing the price tag ever lower.
Like most currently available technologies, Ion Torrent's Personal Genome Machine uses polymerase-based sequencing-by-synthesis chemistry. However, unlike current sequencers, it uses an electronic detection system, rather than cameras or labels, to see which nucleotides are being incorporated into the sequence, and it is based on a semiconductor chip. It senses the pH level changes that occur when nucleotides are incorporated into the sequence. "We load each of the four bases sequentially in a manner that would be analogous to four-channel sequencing from back in the day," Life Tech's Therrien adds. The Ion Torrent machine is priced at just over $50,000 and will launch soon. "We are shipping in volume in November," Therrien says.
Also launching soon — likely in the first quarter of 2011 — is Pacific Biosciences' RS platform. This system, which was unveiled at the Advances in Genome Biology and Technology meeting in Marco Island, Fla., in January, is based on the company's single molecule, real-time analysis sequencing technology, which also relies on polymerase. It uses an array of SMRT cells that each contain about 75,000 zero-mode waveguides, which are nano-scale holes, to which DNA polymerase is immobilized. Cameras then detect the phospholinked nucleotides as they are incorporated into the DNA sequence.
According to PacBio's recent SEC filing — the company aims to raise $230 million for its IPO — the RS platform comes with a touch screen control panel, automated liquid handling, high-performance optics, and a computational blade center and software. The company adds that that the system is to have average read lengths of more than 1,000 base pairs. Sample prep time, using its "SMRTbell sample preparation protocol" is estimated to take as little as "30 minutes of instrument time" and "sequence data is produced in minutes rather than days."
Complete Genomics filed its paperwork to go public about two weeks before PacBio did, and it hopes to raise $86.25 million in its IPO. Complete Genomics takes a different approach to sequencing, relying on patterned DNA nanoball arrays, which it says allows it to "pack DNA very efficiently on a silicon chip." It combines the nanoball technology with its combinatorial probe-anchor ligation read technology to sequence genomes. This approach, the company says, "enables us to read DNA fragments efficiently using small concentrations of low-cost reagents, while retaining extremely high single-read accuracy."
A team from Complete Genomics and its collaborators reported in Science in January that it used this nanoball sequencing approach to sequence three human genomes to an average coverage of 45- to 87-fold each. They reportedly used an average of less than $4,400 in consumables per genome. In addition, a pilot project from a Genentech-Complete Genomics collaboration reported in Nature used Complete's approach to sequence the genomes of non-small cell lung cancer tumor and normal tissue samples to 60-fold and 46-fold depths of coverage, respectively.
In the future
While those companies prepare to launch their products, another wave of technologies and players are following close behind. These new sequencing approaches are drawing on recent advancements in microfluidics and nanopores, and they could drop the price and up the speed of sequencing even more. These technologies and companies are still in their early days, however, and aren't quite ready yet — but they offer a taste of what's to come.
Coming out of David Weitz's group at Harvard University is a microfluidics company, GnuBio, that says it will be able to sequence a human genome for $30 worth of reagents. "The big idea is to just use droplets as test tubes," says Adam Abate, a postdoc in Weitz's lab. The company, with Abate as the PI, was awarded a $240,000 grant as part of this year's $1,000 Genome grants from the National Human Genome Research Institute. "In other words, you would use one droplet to perform one reaction and the advantage to doing this with microfluidics is that you can create and process the droplets at rates of tens of thousands, hundreds of thousands, even millions per second," he says.
Within that droplet, the company is using a variation of the sequencing-by-synthesis method — using ligation rather than synthesis. In it, a probe sequence is hybridized to the DNA and then ligated to another probe, Abate says. A FRET assay is used to show that a ligation reaction has occurred. "The result of that is that you get short snippets of the sequence that you are trying to read," he says, likening it to shotgun sequencing. "Eventually, you'll fill in the entire sequence."
The advantage to this approach, he says, is that it can be done at a high throughput and for a low reagent cost. "If you do a real quick back-of-the-envelope calculation, if you figure you need to do 3 billion reactions in order to sequence a human genome, that's 3 billion times 0.5 picoliter, which is roughly a milliliter of total reagent," Abate says. "You could, in principle, sequence a whole genome for 1 mL of total reagent, and you could do around a million reads per second. That would be about an hour to do a full genome."
Microfluidics, though, has a competitor in nanopores. This approach, too, has the possibility of being a fast and cheap way to do sequencing. Oxford Nanopore is developing a platform based on the work of Hagan Bayley, a professor at the University of Oxford. The company is working on α-hemolysin pores contained within a lipid bilayer for sequencing. Those pores then can be put together as an array of hundreds or thousands of pores within the bilayer, over a set of microwells. The company is working on direct electronic detection of the DNA as it passes through the pore. "In our mind, the whole thing about nanopores is that it allows you to take a direct electronic measurement without using an optical label. The whole benefit of this technology is its scalability: the fact that it's real time, the fact that it's digital, the fact that it is massively scalable," says Zoe McDougall, the communications and marketing director at Oxford Nanopore.
The company is also developing different nanopores for different types of sequencing, though the approaches are built off the same foundation platform, adds Spike Willcocks, vice president of business and corporate development at Oxford Nanopore. The exonuclease-based technology, for which Oxford Nanopore has a commercialization agreement with Illumina, pairs an exonuclease with the nanopore. The exonuclease cleaves the bases from the DNA strand and sends them sequentially into the nanopore to be read.
For the other approach, a whole strand passes through the nano-pore, untouched by an exonuclease. This presents two main challenges: slowing the DNA down as it passes through the nanopore so that the bases can be read, and then recognizing those bases. McDougall adds that recent academic papers have shown progress on those fronts.
Other nanopore sequencing approaches are in even earlier stages of development. Amit Meller at Boston University received one of the NHGRI $1,000 Genome grants to work on his single molecule nanopore-based sequencing, which uses silicon nanopores. To deal with the problem of DNA passing through the nanopore too quickly, Meller's approach converts the DNA to a digital code and then the sequence is read by a CCD camera capturing the light off the attached florescent probe. The new funding from NIH will allow his group to try multiplexing the nanopores. "Our new NIH funding will allow us to realize an instrument … where we have hundreds, if not thousands of nanopores on the same tiny chip. Imagine the chip having a membrane of 100 by 100 micro-meters in area," he says. "You can fit in this area something like 10,000 nanopores and read all those pores silmultaneously on one or two CCD cameras." Much of Meller's work is being commercialized by startup NobleGen Biosciences.
Even further out are graphene nanopores, a material that Delft's Dekker finds interesting for sequencing and is working on in addition to silicon nanopores. Graphene is a single atom thick — which is about as wide as the distance between bases — and it's a conductive material. "In principle, if it gets going with electrical read-out that is label-free, for example, you don't have to add special fluorescent tags. You can just use bare DNA as it is, which would be an advantage," Dekker adds.
This surge toward faster and faster sequencing is simultaneously pushing the cost of that sequencing lower and, many hope, closer to the goal of a $1,000 price tag for a human genome — a milestone that many companies and researchers hope to achieve. It's a crowded race, and just which technology will get there is not yet known.
Of course, it may depend on what is meant by a $1,000 genome. "Does that price include just the costs of reagents, or the amortized cost of the instrument and labor?" asks Oxford Nanopore's Willcocks.
"If you ask the question, what does it cost to sequence a genome, of course you factor in the whole lab component of labor, reagents, capital equipment, but you also have to look at data storage and data processing," adds Life Tech's Therrien.
If you count everything — all of those above costs — Illumina's Flatley says that he thinks the $1,000 price point will be reached within "three to five years."
In addition, getting there first doesn't necessarily mean that a given technology will become king of the hill. Abate notes that GnuBio's approach might be best suited to certain applications, while others would be better able to pick up, say, tandem repeats. "Because of the nature of sequencing, there is going to be room for more than one technology," Abate says.