For ABI, Developing Agencourt's Sequencing Technology Is a High Priority
NEW YORK (GenomeWeb News) - Applied Biosystems last week said it wants to offer early-access DNA sequencing services based on Agencourt Personal Genomics' new platform by the end of the year as the company looks to gain a foothold in the next-generation sequencing technology market.
Last week, ABI said it will acquire the Agencourt Bioscience subsidiary from Beckman Coulter and its other shareholders for approximately $120 million in cash. The deal comes around eight months after the Applera unit bought an undisclosed stake in sequencing startup VisiGen Biotechnologies and is the latest step in ABI's ongoing strategy of investing in different kinds of next-generation sequencing platforms.
With APG, ABI will acquire a near-term sequencing technology to address the "$100,000 genome." APG's platform, a sequencing-by-ligation technology partly based on the method published by George Church's group in Science last fall, is scheduled to appear in early access labs in 2007 and is therefore behind its top rivals: 454 Life Sciences has been selling its GS20 sequencer for over a year, and Solexa launched its early-access instrument at a conference last week and wants to ship the first units by the end of this month.
But based on ABI's track record of developing and commercializing DNA sequencers, APG, whose parent Agencourt Bioscience has had a long-standing commercial relationship with ABI, might still be able to gain a place in the market.
ABI evaluated more than 40 sequencing technologies both from private companies and from academic labs before settling on APG. According to Kevin Corcoran, vice president and general manager of ABI's genetic analysis business, the company assessed these platforms by comparing total throughput, scalability, data quality, read length, cost per finished base, number of possible applications, and commercial feasibility.
Some technologies dropped out because they were not yet developed far enough. "If it looks great but it does require three miracles to be pulled off before you can use that technology, that's something you want to stay away from," Corcoran said.
After ranking the technologies, APG came out on top. "We felt that it scored the highest in most of these particular areas," Corcoran said.
APG was also asked to sequence a BAC clone of human DNA of which around one-third comprised repeat units. The company performed this task and returned data that were around 99.9-percent accurate.
"We were quite impressed, given the complexity of this particular BAC, that they were able to turn it around quickly and with such accuracy," Corcoran said. He and his team also visited APG's Beverly, Mass.-based labs and "were impressed by the basic ease of use in getting this process done," he added.
Another potential advantage APG had was the longstanding relationship its parent has had with ABI: Agencourt Bioscience's sequencing center, which provides sequencing, SNP discovery, and library construction services, owns 31 ABI 3730XL sequencers. The sequencing service business, which is not part of APG, will remain with Beckman.
"Agencourt Bioscience is a very good customer of ours," said Corcoran, and ABI has been aware of APG's work on next-generation sequencing for a long time
When Beckman acquired Agencourt Bioscience last year for up to $140 million in cash it only took a 49-percent stake in the still-nascent APG unit, which was incorporated in January 2005. At the time, "we couldn't really agree on a valuation for APG because it was very early stage," said Kevin McKernan, APG's president and formerly Agencourt Bioscience's co-CSO. So "we just cut it aside and cut it down the middle."
As a result, Agencourt Bioscience's shareholders, which included founders and management, the Whitehead Institute for Biomedical Research, MIT, and Packard BioScience (now PerkinElmer), remained in control of APG's board.
According to McKernan, though Beckman "always kept an interest" in APG and "gave excellent advice in terms of how we should develop" its sequencing technology, the company realized it could not ensure developing an instrument within the shrinking timeframe dictated by the mounting competition from 454, Solexa, and others.
When Agencourt founded APG, "the time-sensitivity of getting [a sequencer] to market was not as evident as it has become today," McKernan told GenomeWeb News last week. Since then, getting to market within the next year or two has become crucial, "otherwise we will miss the window," he added.
ABI, on the other hand, is "much more vested in [the DNA sequencing] marketplace" and could make APG's instrument a high priority, McKernan said. So when ABI came knocking, Beckman agreed to sell its APG stake for around $50 million. The remaining $70 million will go to APG's other shareholders. The deal is expected to close in the third quarter.
APG's technology could benefit from being developed at ABI because the company has a history of coming from behind to capture winning market shares. For instance, back in the late 1990's, PerkinElmer's PE Biosystems division -- which included Applied Biosystems -- was racing behind Amersham Biosciences to get its high-end capillary electrophoresis sequencer to market. At the time, the ABI 3700 sequencer was 18 months behind Amersham's MegaBace 1000 instrument, according to McKernan, but it "still took the market by storm," he said. That history "certainly resonated with people here, that these folks know how to build instruments quickly and get them to market," McKernan said.
The APG Platform
So what does APG's technology offer? It currently produces two paired 25-base reads per template and close to 100 million bases per run, which takes about 3 days, ABI's Corcoran said. According to McKernan, longer reads of at least 50 bases are possible in principle, but "once you get to 25 you have met most of the applications' needs." In terms of throughput, ABI's goal will be to ramp up the raw data output to 1.2 gigabases, or 1.2 billion bases, per run, according to Corcoran.
APG's system offers features that will distinguish it from some of its rivals. For example, APG's technology does not have a problem with homopolymers, Corcoran said. Solexa has claimed in the past that its system also offers good homopolymer accuracy, compared to 454's system. In addition, APG's libray construction strategy allows for paired-end sequencing, a feature that both 454 and Solexa are currently working on but do not yet offer.
Also, APG's technology amplifies the template in a single tube, facilitating its preparation. And APG's use of ligase instead of polymerase, which is used by sequencing-by-synthesis approaches like Solexa's and 454's, leads to higher fidelity, Corcoran claimed.
APG's sequencing-by-ligation strategy is based in principle on the method published by George Church's group last fall but has evolved to increase read length and accuracy, according to McKernan. APG holds an exclusive license to the technology from Harvard for certain fields of use. In addition, the company holds a number of licenses to other outside patents, "mostly" from academic groups, and has about a dozen patents or patent applications of its own on aspects of the method according to McKernan.
To sequence DNA, APG scientists first construct cell-free libraries of genomic DNA using either a strategy similar to Church's, which creates paired-end sequences, or one that ligates priming sequences to the ends of genomic DNA fragments. Next, they amplify the target DNA by emulsion PCR and enrich it using capture microbeads, which are immobilized on a gel-covered glass slide. APG then adds primers and fluorescently labeled ligation probes and uses an epifluorescence microscope to reads out the results.
Unlike Church's method, which cannot perform more than one ligation in each direction, APG's method can do several sequential ligation reactions, leading to longer reads. This ligation chemistry -- which is patent-protected and involves a silver-cleavable phosphorothiolate linkage in the ligation probes -- is APG's most important enhancement of the technology.
"This patent estate has a fair amount of novel tricks in it for not only doing ligation sequencing but ... also an error checking procedure ... that's quite helpful in improving the accuracy," McKernan said.
"We think that this technology is going to be able to deliver on the '$100,000 [human] genome,'" Corcoran said, a price that includes all project costs. The fold-coverage required will be dictated by the technology, he said, and has not yet been determined. Solexa has said it wants to sequence a human genome at 20-30x coverage for $100,000 or less by the end of this year.
So what is ABI's strategy for commercializing APG's technology? The company plans to start offering early access sequencing services for "key individuals" by the end of the year, followed by early-access instrument placements sometime in 2007, according to Corcoran. After that, ABI plans to enter the market with a "very robust instrument and reagents."
The price for the instrument will be "competitive" with ABI's 3730XL capillary electrophoresis sequencer, the high-end version of which lists for $365,000.
ABI will market the instrument for different applications, the two most important ones being high-throughput sequencing and gene expression analysis, according to Corcoran.
For gene expression studies, the technology could be especially useful for researchers wishing to profile the transcriptomes of organisms for which no micorarray is available, such as in agriculture, he said. In addition, measuring gene expression by sequencing is more sensitive and reproducible than microarray data, he claimed, which would make it interesting for researchers looking at lowly expressed genes, such as neuroscientists.
Another application will be high-throughput genotyping, especially if the aim is to discover novel gene differences, Corcoran said.
On the bioinformatics side, ABI said it plans to integrate the software to analyze data from its CE sequencing instruments and from APG's platform. "Integrating those two datasets is going to be ultimately what gives people the highest level of performance," Corcoran said.
In addition, the company plans to make the software for the new instrument an open system "from day one" so that third parties can add to it, Corcoran said. For example, he said, users of existing ABI sequencers have long been asking for access to the file formats and the company is "actively working on" opening the software of its CE sequencers, he said.
ABI said it plans to retain APG's group of about 20 scientists, who will remain in the company's Beverly headquarters and closely collaborate with the ABI R&D team in Foster City, Calif. More staff would be added "as needed" on either coast, according to Corcoran. "It doesn't hurt having things in two different time zones to sort of swing-shift the R&D," McKernan said, adding that the teams would "get three more hours out of our days."
But the APG acquisition is not ABI's last word in its pursuit of next-generation sequencing technologies. The company is doing internal research and looking to invest externally into so-called real-time single-molecule sequencing technologies that produce read lengths in excess of 500 bases.
"That drove our investment in VisiGen" last year, Corcoran said, a platform ABI is still very interested in. These technologies -- which include, for example, nanopore sequencing and zero-mode waveguide sequencing, according to Corcoran -- are "probably seven to eight years out," he added