As next-generation sequencer vendors work to improve the chemistry, hardware, and software for their instruments, other companies are developing front-end technologies that promise to simplify sample prep or allow researchers to select genome regions and multiplex thousands of samples.
Less than two weeks ago, Population Genetics Technologies, a startup based in Cambridge, UK, said it has raised £3.8 million (approximately $7.5 million) in a first round of venture capital funding to develop a technology that can tag individual samples and select DNA regions from entire populations for multiplexed sequencing applications, and said it could lower the cost of sequencing in such projects. Separately, last week, Dublin, Calif.-based Microchip Biotechnologies said that it is collaborating with researchers at Stanford University to develop a sample-prep system for high-throughput pyrosequencing that it says will reduce the time, labor, and reagent costs of the process.
PGT is developing technology invented by Nobel Laureate Sydney Brenner. According to newly hired CEO Mel Kronick, this has two parts: a tagging technology that allows users to identify individual samples in a mixture, and a sorting technology that enables them to select particular DNA regions of interest, he told In Sequence. Because of its early development stage, the company is not yet disclosing details of its technology.
According to a press release from the Wellcome Trust, the technology uses “two ‘genetic bar-codes:’ one to identify the individual genome from which a fragment originated, and the other to identify that fragment’s position in the genome.”
Instrument vendors as well as their customers have already developed a number of barcoding methods to sequence dozens to hundreds of samples in parallel (see In Sequence 8/7/2007). Most recently, a group led by Rob Knight at the University of Colorado developed an error-correcting barcoding technology that allowed his team to sequence 1,544 samples in parallel on 454’s platform, according to a study in this month’s Nature Methods.
Also, researchers and companies such as NimbleGen Systems and Agilent Technologies have been working on ‘genome partitioning’ technologies that allow them to select specific portions of the genome for sequencing, such as exons, using microarrays or solution-based oligonucleotide libraries (see In Sequence 11/6/2007).
PGT’s selection technology differs from these, according to Kronick. “There are aspects to our selecting process that are more sophisticated and nuanced than traditional ‘capture’ methods,” he said. Moreover, it is possible that PGT will combine its tagging technology with existing capture methods. “We can envision using our technology to do both [tagging and selecting], or we may want to integrate our workflow with other ‘selecting’ technologies that will complement our offerings,” Kronick said.
Overall, the company is designing the technology to be scalable so it can be applied to large numbers of samples. “We realized that when people are doing population studies, the numbers get up into the thousands,” Kronick said. “We are designing the system with those kinds of numbers in mind.”
Users, in principle, can freely determine which part of the genome to select for sequencing. “They are not restricted to particular sections of the genome,” Kronick said, cautioning that the company still needs to validate this aspect of its technology.
In addition, he said, the company is “trying to implement the processes in a way that is going to be easy for the user, and cost-effective,” he said, adding that he believes the technology will be cheaper than existing methods.
PGT was founded by Brenner; Sam Eletr, a co-founder of Applied Biosystems as well as the former chairman of Lynx Therapeutics and a former director of Solexa; and Philip Goelet, former CEO of Molecular Tool, now called Orchid Biosciences. The three founded Compass Genetics, a partnership that licensed the technology to PGT.
In 2005, the company received £1.1 million (about $2.2 million) from the Wellcome Trust to prove the feasibility of the technology, which allowed the company to establish operations at the Babraham Research Campus, an incubator a few miles south of Cambridge that houses more than 20 biotechnology startups.
PGT received its first US patent, No. 7,217,522 entitled “Genetic analysis by sequence-specific sorting,” last May and has various other patents pending, according to Kronick.
“We need to make the whole process easier, and automate it more, to make it usable in larger communities than large sequencing facilities.”
Last month, the company announced that it had received £3.8 million in a Series A private-equity funding round from Auriga Partners, Noble Fund Managers, and Compass Genetics. PGT will use that funding to hire additional staff that will enable it to “take the concepts that we have been reducing to practice over the last two years” and develop them further.
“Our focus, now, is taking the technology to the point where its path to product becomes very clear,” Kronick said.
The plan, at the moment, is to “position the technology as generically as we can,” so it could be implemented, with modifications, as a front-end to different next-generation sequencing platforms, he said. “We have as a goal to make it as applicable to many different platforms as possible.”
Company scientists are also designing the elements of the technology in such a way that they can be easily automated. “We want to keep our options open at this point,” Kronick said, referring to the company’s automation plans. However, he mentioned as possible partners companies like Raindance Technologies, Fluidigm, Microchip Biotechnologies, as well as conventional liquid-handling robotics companies.
PGT has not decided yet whether it will acquire its own next-generation sequencer or rely on outside service providers as it currently does. “It has to do with cost, convenience, [and] what our [sequencing] demand will be,” Kronick said. “These machines are not inexpensive.”
The company is unlikely to present its technology in public for at least another 12 months, he said, although it might engage in “some very limited early-access programs” earlier than that.
“We have the basic strategy pulled together in a way that I am pretty confident we are going to be able to achieve our goals,” Kronick said. “But I am also enough of a scientist to know that the proof is in the pudding, that we have to show the data that validates the fact that we can achieve those kinds of things.”
Prepping for Pyrosequencing
Microchip Biotechnologies, on the other hand, said last week that it is collaborating with Mostafa Ronaghi, a senior research associate at the Stanford University Genome Technology Center, to develop a microfluidic-based sample-preparation system for high-throughput pyrosequencing.
Ronaghi’s group has been working on a miniaturized next-gen pyrosequencer (see In Sequence's sister publication, GenomeWeb Daily News, 1/2/2007), “but we need to have an automated sample preparation,” he told In Sequence last week. “We need to make the whole process easier and automate it more to make it usable in larger communities than large sequencing facilities.”
Under a joint NIH grant with Ronaghi, MBI, which is based in Dublin, Calif., has already developed different modules for the process, which it is currently integrating into a single chip. The partners are also waiting to hear about a pending grant they have applied for.
Except for fragmenting genomic DNA, the chip will take on all steps of sample preparation, Ronaghi said, including setting up and running the emulsion PCR, breaking the emulsion, capturing and enriching the amplified fragments, and purifying the DNA.
Depending on the throughput, the device could reduce the sample prep time from currently two days to between 12 hours and 20 hours, according to Ronaghi. Unlike the current process, it would not require manual steps.
MBI has already developed an improved method for amplifying long DNA fragments on beads, a prerequisite for long sequence reads, Ronaghi said. The reason that process had been inefficient before, he said, was that the bubbles in the emulsion were not of uniform size, and beads captured in small bubbles did not amplify well.
MBI developed a chip-based process that generates a “monosized” emulsion, he said, and has been able to amplify fragments longer than one kilobase. “Having those long fragments means that we can extend the read length of pyrosequencing beyond the 400 base pairs that 454 has announced recently,” he said.
MBI researchers have also shown that the emulsion PCR works on a chip and that they can capture beads, purify amplified DNA fragments, and transport them on-chip. “The next year, they will spend time on integrating and optimizing the whole process,” Ronaghi said.
According to MBI CEO Stevan Jovanovich, the company plans to start selling the device two years from now at the earliest, “and given commercial realities, it would need to happen by three years,” he told In Sequence.
At this point, MBI has not decided whether the device will be sold for a specific sequencing platform, or for several systems. “It comes down to commercial decision and potential agreements with the different backend platforms,” Jovanovich said.
“You could easily use it for a 454 system,” according to Ronaghi, and although it will be optimized for pyrosequencing, “with a small modification,” it could also be used for ABI’s SOLiD sequencer, which like 454 uses emulsion PCR as part of the sample preparation.
The device could even be used for the early steps of Illumina’s sample prep for the Genome Analyzer, “but since their sample prep diverges quickly from the others, that would be a different project,” Jovanovich said.
Ronaghi intends to combine the device with the pyrosequencer he has been developing at Stanford. He and his colleagues have already come up with a detection chip that is “very scaleable” and has “lots of sensitivity.” Recently, they have integrated the different parts of the system into a “small box” and are currently “getting ready for large genome sequencing applications,” he said, adding that Stanford has not yet licensed the technology to a company.