NEW YORK (GenomeWeb News) - As second-generation DNA sequencers become available, the third generation is waiting in the wings.
One of these is a nanopore array-based optical readout system for single DNA molecules developed by Boston University physicist Amit Meller. The technology, which will read binary-encoded DNA prepared by Norwegian startup LingVitae, is “very close” to undergoing a proof-of-concept study, according to Meller.
Eventually the system, which Meller has been developing with a two-year 2004 NHGRI Advanced Sequencing Technology grant, will be “orders of magnitude faster than all the others” currently on or soon to hit the market, he said.
Meller said he initially tried to read DNA by measuring the ionic current as the molecule travels through the nanopore, but “we had a lot of different problems.” For example, the current was averaged over a number of bases, and the DNA moved much too fast to discriminate between bases, “which lowered the excitement about single-molecule sequencing in this way,” he said.
That changed when a co-worker of Preben Lexow, LingVitae’s CEO, met with Meller at a conference in Japan in 2001. Having heard about Meller’s difficulties, the co-worker suggested combining the DNA-preparation technology LingVitae had been developing with Meller’s nanopore approach, and the two groups began to collaborate.
The technology, invented by Lexow, biochemically converts short fragments of target DNA into binary-encoded DNA molecules in a three-step enzymatic process that is repeated several times.
The method uses two types of short, pre-synthesized oligos, or code units, different combinations of which are assigned to represent a base in the original DNA. For example, using this binary code, an “A” could be represented by oligo 1 followed by oligo 2; a “C” by two copies of oligo 1; etcetera.
In the first step, the first two bases of the target DNA are cleaved off with a restriction enzyme. In the second step, they are identified with a ligase-dependent recognition system. Finally, they are replaced with four code units that are attached to the other end of the target molecule.
The conversion results in so-called “design polymers” that can be read by a nanopore-based platform such as Meller’s. All the platform needs to do is distinguish the code units, typically 10 bases in length, instead of reading every single base in the original DNA.
To do this, Meller hybridizes fluorescently labeled oligos that are complementary to the code units to the design polymer. As the DNA passes through the nanopore, which is too narrow for double-stranded DNA, the oligos peel off and can be read with a CCD camera.
“This solves the main issues associated with nanopore sequencing: We only need to resolve chunks of 10 bases instead of single bases, and the DNA moves much slower due to the peeling-off process,” Meller said.
Though LingVitae’s DNA preparation appears complex, any single-molecule sequencing technology requires the DNA to be prepared in some way in order to remove nucleosomes, break it into manageable pieces, or make it single-stranded, according to Meller. “Adding the complexity that LingVitae brings is not a major departure,” he said.
Another advantage of the approach is that it can be parallelized, he added. Meller has already produced arrays of 36 nanofabricated pores in a six-by-six design and hopes to get to 10,000 in the future.
”This way we can boost the throughput of this technique by orders of magnitude,” he said. “It will give us a sequencing rate of millions of bases per second” on a 200 micron square chip holding the array.
According to Meller, using 10,000 nanopores in parallel instead of one also reduces the likelihood that contaminants such as proteins or lipids will stop the process. “We learned from our experience that if you have only one precious pore which you built, and then you clog it, you are doomed,” he said.
In addition, sequencing data can be analyzed in real-time as it is collected, making error correction faster than with other platforms that require a full run to be finished before the sequence of each fragment is known, said Meller.
So far, he has been able to show that he can use nanopores to strip off the oligos, and that he can produce arrays of pores. What is missing still is a proof-of-concept of being able to read out the sequence of displaced oligos.
However, he thinks “that we will show the feasibility in a matter of months,” and added that several companies have already approached him about possible commercialization. He declined to elaborate.
Though Meller’s platform is still far from being a commercial product, in principle his approach competes with other single-molecule technologies, such as Helicos’ polymerase-based platform.
Meller said he believes he is one step ahead of many other nanopore-based technologies that still need to prove a lot of basic concepts. “We have all the components in hand, yet we need to do more research to conclude that we can achieve what we plan to,” he said.
In general, nanopore sequencing and other single-molecule approaches promise to “eliminate some of the concerns about reaction phasing, fidelity of PCR, etcetera, that plague non-single molecule approaches” such as 454’s and Solexa’s, according to Elaine Mardis, co-director of the genome sequencing center at Washington University School of Medicine.
However, nanopores pose a number of challenges, such as being able to detect and identify each nucleotide, and making DNA with secondary structure enter them, she told GenomeWeb News in an e-mail.
Even if Meller’s readout platform works as planned, the scientists have to prove that the DNA-prep technology does not introduce mistakes into the sequence. “They first would have to show data demonstrating an extraordinary level of fidelity in transitioning the input DNA into the ‘new’ binary-encoded DNA molecule,” Mardis cautioned.
LingVitae’s Lexow would not disclose the current accuracy of the conversion “as the product development is still in its early stage,” he said. To date, LingVitae has been able to convert 20mers of DNA, which represents the read length of the technology, and hopes to ramp this up to 24 by next year, and later on to even longer pieces. The current read length “is a limitation when it comes to some sequencing, especially de novo sequencing, or large repeats,” Meller conceded.
At the moment, LingVitae’s technology can convert approximately 100 gigabases of DNA per sample. Lexow said he plans to miniaturize the reactions and move them to a disk-based format by next year, which would cut the cost of consumables. “We expect to be very price-competitive” in terms of consumables compared to second-generation sequencing platforms, he said.
Up until now, LingVitae, has been working on applications for its binary preparation technology with a range of academic partners. Beside Meller partners include Aaron Bensimon at the Institut Pasteur in Paris, Christoph Gerber at the University of Basel in Switzerland, and a number of scientists based in Oslo.
In general, the company believes its technology will enable third-generation platforms like Meller’s to sequence DNA without having to reach single-base resolution.
But it’s still a young company: LingVitae has been operating in stealth mode over the past few years but launched its website [LINK: http://www.lingvitae.com/] yesterday. It has been developing the DNA-preparation method since it was founded and plans by mid-2007 to offer the technology to commercial OEM partners that would develop it for their own readout platforms, not necessarily nanopore-based ones.
The exact nature of the code units will be dictated by the partners, Lexow said. “All readout companies will probably have their own preference when it comes to design polymers.”
Lexow founded LingVitae in 2002 after his first DNA-sequencing venture, Complete Genomics, faltered earlier that year. That company, which developed a DNA-preparation technology similar in concept to LingVitae’s, was to offer DNA sequencing as a service, he said.
LingVitae currently runs on a mix of public grants and private equity and has burned through $5.6 million so far, according to Lexow, who would not disclose the total funding for the company.
Julia Karow covers the next-generation genome-sequencing market for GenomeWeb News. E-mail her at [email protected].