The National Human Genome Research Institute last week announced $15 million in new research grants to 11 investigators for developing new, cheaper DNA sequencing technologies.
The grants, awarded under NHGRI’s Advanced Sequencing Technology program, fund three groups working on near-term technologies for the so-called $100,000 genome, and eight groups focused on “revolutionary technologies” for the $1,000 genome (See details of the projects below).
Though the institute was not explicitly looking for any particular approaches this year, the funding portfolio has shifted somewhat from near-term to longer-term technologies since the program started in 2004.
“There are still certainly problems to be solved for the $100,000 genome” — both technological issues and commercialization challenges — according to Jeff Schloss, NHGRI’s program director for technology development. “But I think the emphasis for an agency like NHGRI, [which is focused on] research questions, is shifting now toward the next generation of technology,” Schloss told In Sequence last week.
He said the institute uses the $100,000 and $1,000 numbers “almost as shorthand” for the actual aim of the program: reducing the cost of sequencing by either two or four orders or magnitude, compared to 2004 prices.
“What [we] really mean is, for whatever data quality, [or] kind of sequence data, you are talking about, take whatever it cost to do that in 2004 and reduce it 100-fold or 10,000-fold,” Schloss said.
For example, it cost about $10 million in 2004 to generate a high-quality draft of the human genome, hence the $100,000 and $1,000 numbers. Schloss said he does not believe any of the currently available next-generation technologies has reached the $100,000 genome target yet, at least not by NHGRI standards.
As a quality standard, NHGRI is using the mouse genome assembly that was published in 2002, which Schloss described as a “high-quality draft with reasonable contiguity.” Regarding the costs for achieving this data quality, “we mean everything,” he said. For example, costs should include sample acquisition, DNA preparation, sequencing, and genome assembly, and comprise reagents, personnel, electricity, and instrument depreciation.
So far, NHGRI has not tried to calculate the cost for any of the new technologies that have been funded under the Advanced Sequencing Technology program “because they are changing so fast, it would be impossible to do it,” Schloss said. However, the institute said in a press release last week that it “still costs as much as $5 million to sequence 3 billion base pairs,” or a human genome, not specifying any technology. That is a very rough estimate, according to Schloss. “The point is that we have got a long way to go before we have a really high-quality sequence for $100,000,” he said.
It is possible that this bar is actually higher than what will be required for medical sequencing, he admitted, but “rather than pick the lowest common denominator, we set a relatively high target.”
NHGRI hopes that in two to three years, additional $100,000 genome technologies funded under the program will be moving into commercialization, or be licensed to other groups.
For the $1,000 genome technologies, the institute hopes they will be “on a path toward commercialization” after 10 years. But it is hard to predict how long it might take to develop these long-term technologies. “Any time you say something is a 10-year project, that means you don’t know how long it’s going to take,” Schloss said.
Nanopore-based approaches in particular still face “huge science challenges,” Schloss said. Though the ability to manufacture pores and attach sensors to them has improved, scientists still cannot use them to distinguish the four bases in sequencing. But a few groups, he said, are “getting very close” to using modified protein nanopores to distinguish the bases. The idea of nanopore sequencing is attractive, because it does not require any reagents other than those used to purify the DNA. “Whether it’s practical, we don’t know yet,” Schloss said.
“The point is that we have got a long way to go before we have a really high-quality sequence for $100,000.”
The good news is that “nobody is hitting any roadblocks, as far as we can tell, so the science seems to be moving forward,” he said. “At the very least, no one has demonstrated that nanopores cannot work.”
The $1,000 genome target could also be achieved by reducing the amount of reagents used by the current $100,000 genome technologies. That, Schloss believes, will require “some big engineering changes” such as miniaturization or process improvements.
So far, NHGRI is happy with the outcome of the program. “Clearly, we supported several of the technologies that are either out there now or are about to be out there,” Schloss said. Among these are 454 Life Sciences’ Genome Sequencer, which is already being used in production sequencing at genome centers; ABI’s SOLiD sequencer, which will hit the market this fall; and Helicos BioSciences’ HeliScope, which is scheduled for commercial release later this year (see related story in this issue). Illumina’s Genome Analyzer was not developed with NHGRI funding.
The new awards add to $83 million NHGRI has already committed to the Advanced Sequencing Technology program since it kicked off in 2004: In its first year, the institute awarded $32 million in grants; in 2005, $32 million; and last year, $13 million. But because some of the earlier grants are still ongoing, the program’s annual funding level has actually remained the same, at about $25 million, according to Schloss. That funding level will probably stay the same for the next several years, he said.
A Look at the Projects
The following research groups received funding under the program last week. For more information, click here.
The ‘$1,000 Genome’ Grants
Richard Fair at Duke University will receive $3.7 million over the next three years to work on “continuous sequencing-by-synthesis, based on a digital microfluidic platform.” His group already obtained a two-year, $510,000 feasibility grant under the program in 2005. In that project, he and his colleagues demonstrated that pyrosequencing is possible on the platform, Fair told In Sequence by e-mail this week. The team also figured out how to immobilize DNA so that nucleotides can be presented sequentially, and how to wash the DNA between each sequence step.
The technology will be able to separate the incorporation and detection steps, Fair said, so that detection is not limiting the throughput. The researchers can also sequence through homopolymer regions by continuing to provide nucleotides. “Unlike the 454 platform, we are no longer constrained to deliver the same nucleotide concentration to all wells simultaneously, and we can wait instead for the light signal to be detected from an individual well before deciding whether to add the next nucleotide or add more of the same nucleotide,” Fair said. “In other respects, our approach is complementary to 454.” The scientists still need to develop ways to provide nucleotides in large numbers of droplets for continuous sequencing, with a 350-base sequence requiring almost 2,000 droplets. They also have to further optimize reagents and chemistries.
Advanced Liquid Logic, one of Fair’s collaborators, plans to develop and commercialize the platform. The other collaborators are Ron Davis, Peter Griffin, and Mostafa Ronaghi from Stanford University.
Stuart Lindsay at Arizona State University won a three-year, $877,000 feasibility grant to develop molecular wires that are flexible and sensitive enough for sequencing. The method relies on detection by electron tunneling, Lindsay said in an e-mail message. What is new, he said, is the incorporation of chemical recognition into the tunneling approach. His group has already developed a recognition reagent that “works with the bases to ‘read’ DNA sequence,” he wrote.
Longer term, the technology needs to become part of a small enough device to access a single base at a time, he said. The most important improvement over current technologies might be in read length. Lindsay estimates that “it may be possible to read more than 80,000 bases at about 1 millisecond dwell time per base,” and it “should not be very difficult to increase this number substantially.”
Lindsay’s collaborators are Peiming Zhang, a DNA chemist, and Otto Sankey, a theoretical physicist, at ASU. Greg Timp at the University of Illinois-Urbana Champaign has advised the group on nanopore technology. Lindsay is also working with Trevor Thornton and Hung Chang of the Arizona Institute for Nano-Electronics on device design. A commercial version of the technology is at least eight years away, according to Lindsay, since “the nanofabrication aspects make this very unpredictable.”
In 2004, Lindsay received another feasibility grant under the NHGRI program to evaluate a “molecular reading head for single-molecule DNA sequencing.” However, that technology “turns out to be just on the wrong side of feasible” for sequencing applications, he said.
Xinsheng Sean Ling of Brown University received a three-year, $820,000 feasibility grant for “hybridization-assisted nanopore DNA sequencing.” Ling’s group has already shown that it can control the speed of DNA translocation, he told In Sequence by e-mail, but there are still some surface chemistry issues to be solved concerning the performance of silicon-based nanopores.
“The pores are still too noisy to my liking,” Ling said. Long term, the technology should outperform current sequencing-by-synthesis approaches in read length and throughput, he said. Cost will depend on how the researchers will scale up the technology, and a commercial version is five to 10 years away, according to Ling.
Ling’s group is collaborating with NABsys, which he co-founded in 2004. NABsys won its own two-year, $480,000 SBIR grant under the program, the only grant to a company this year. The Providence, RI-based startup has been working on marrying sequencing-by-hybridization and nanopore sequencing (see In Sequence 1/9/2007).
“The next big milestone at NABsys is achieving reliable detection of probes bound to target DNA to within the positional error allowed by the reconstruction algorithm,” John Oliver, the company’s vice president research, told In Sequence by e-mail. He and Amit Basu at Brown University will work on the platform’s chemistry, while Ling will refine the nanofabrication and work on controlling translocation. Franco Preparata and Eli Upfal at Brown will be developing the reconstruction algorithms, according to Oliver.
Since January, the company closed on $750,000 of a $1 million limited equity round, Oliver said. It also increased its headcount from three to four and plans to add more personnel “in the near future.”
Wlodek Mandecki from the University of Medicine and Dentistry of New Jersey will use his three-year $1.67 million award to develop a “ribosome-based single-molecule method to acquire sequence data from genomes.”
The main problems to be solved relate to gathering fluorescence signals from single ribosomes translating an mRNA molecule, Mandecki told In Sequence by e-mail. Rather than using DNA polymerase or ligase, this method uses the ribosome “as the engine to collect the sequence information,” he said.
The sequencing rate should approach the speed of the movement of the ribosome along the mRNA, which is several tens of nucleotides per second, and the read length will be the length of the mRNA, or several hundred nucleotides. No PCR amplification is involved in the method, which Mandecki wants to parallelize to a high degree.
Mandecki, who is also president and chief scientific officer of PharmaSeq, a Princeton, NJ-based DNA technology company, is collaborating with Emanuel Goldman and Hieronim Jakubowski from UMDNJ; Barry Cooperman and Yale Goldman from the University of Pennsylvania; and Julian Borejdo, Ignacy Gryczynski, and Zygmut Gryczynski from the University of North Texas Health Science Center.
Andre Marziali from the University of British Columbia in Vancouver obtained a three-year, $746,000 grant to work on a nanopore array force spectroscopy chip for rapid clinical genotyping. His group is using force spectroscopy rather then direct optical or electronic measurements.
“This method precludes us from collecting enough information to read long stretches of DNA sequence,” Marziali explained in an e-mail. As a result, he is developing a rapid genotyping rather than a sequencing technology. However, “much of the development we will be doing, specifically in nanopore fabrication, will be useful to the group that are aiming for direct DNA sequencing,” he said.
Under a 2004 feasibility grant from the same NHGRI program, his group was able to demonstrate that the method is capable of resolving differences in the target DNA that are as small as single nucleotide changes. “Most of the problems to be solved at this stage are in porting the method from organic pores ... to synthetic pores that are more amenable to commercial instrumentation development,” he said. Marziali plans to seek a commercial partner for the project in two to three years.
Robert Riehn from North Carolina State University has won a two-year, $439,000 feasibility grant to explore whether lateral electronic transport through DNA can be used to obtain sequence information.
“We hope to ultimately achieve superlong read lengths, up to megabase pairs, at competitive cost and with vastly reduced need for assembly,” Riehn said in an e-mail message. He also plans to run many devices in parallel.
Similar to David Schwartz from the University of Wisconsin at Madison (see below), Riehn will use nanochannels to stretch out the DNA. However, while Schwartz immobilizes the molecules, “in our technique, the molecules can fluctuate, and thus the error is reduced further and further as the same molecule is observed for longer and longer times,” Riehn said.
Riehn is hoping to demonstrate single-base-pair resolution on moderately long DNA within three to five years and to commercialize the technology within seven to 10 years.
Kumar Wickramasinghe at the University of California, Irvine, will receive $2.18 million over three years to develop high-throughput, low-cost sequencing using probe tip arrays.
Integrating optical detection with the ultrafast electrophoresis system the lab has developed “is one of the major” problems he still needs to solve, Wickramasinghe said in an e-mail message. He is hoping to achieve 100 base pair reads with similar accuracy as standard Sanger sequencing, but his method will be several orders of magnitude faster. It will also be two orders of magnitude cheaper than current next-generation technologies, he said.
Wickramasinghe is collaborating with Bob Moyzis at his university, who “will do all the biochemistry for the project.” According to his estimate, it will take about six years to commercialize the technology.
The “$100,000 Genome” Awards
Jeremy Edwards from the University of New Mexico School of Medicine in Albuquerque received one of the four $100,000 genome awards, a three-year, $900,000 feasibility grant to sequence the human genome by polony sequencing. Edwards is working closely with George Church’s group at Harvard on polony sequencing and shares technical improvements with them (see In Sequence 6/12/2007).
However, “there are several specific differences in the technology and we are working to address different scientific questions,” Edwards told In Sequence by e-mail. Though he has no immediate plans for commercialization, “I have several paths that are not public knowledge at this time,” he said.
Two awards under the $100,000 genome program went to Jingyue Ju at Columbia University for two separate projects.
Under a two-year, $644,000 feasibility project, he will generate a library of reversible terminators for pyrosequencing, in collaboration with Mostafa Ronaghi from the Stanford Genome Technology Center. The researchers have already generated some feasibility pyrosequencing data using reversible terminators, which they have submitted for publication, Ju said in an e-mail. The main challenge is to increase the read length by optimizing the nucleotide analog synthesis and to automate the process, he said.
That technology “will address the issue with homopolymeric runs in pyrosequencing,” Ronaghi said in an e-mail message. “If that is successful, then it can be implemented in 454’s pyrosequencing protocol and Richard Fair’s sequencing scheme.”
Under his two-year, $2.22 million grant, Ju will work on an “integrated system for DNA sequencing by synthesis.” This is a continuation of a project for which he won a grant in 2004 under the same program. Ju and his colleagues will now focus on increasing the read length for the SBS approach using fluorescent nucleotide reversible terminators, he said by e-mail. Last year, Columbia University licensed Ju’s SBS chemistry exclusively to Intelligent Bio-Systems, which is developing it into a commercial sequencing platform (see In Sequence 2/1/2007).
Ju also has an ongoing three-year $970,000 grant, entitled “Modulating Nucleotide Size in DNA for Detection by Nanopore,” that he won under the program in 2005.
David Schwartz from the University of Wisconsin, Madison, won a three-year, $882,000 feasibility grant to work on sequence acquisition from mapped single DNA molecules. He is developing a “new scheme” that is “designed to obtain sequence information from large genomic DNA molecules, using developmental cues from the “old” sequencing approaches that we published several years ago,” he said in an e-mail.
Schwartz’s UWA collaborator Michael Newton will contribute to the statistical analysis of the data.
Schwartz said he would either like to see the new system used with existing or emerging next-generation sequencing platforms or as a standalone system. “[It’s] hard to put a timeline on this, but sooner is better,” he said.