'Shotgun Wedding': Public, Private Drosophila Sequencing Agreement Should Speed Project, Ensure Accuracy

When J. Craig Venter proposed last May to use the "shotgun" technique to sequence Drosophila, many scientists doubted that blasting such a large genome into billions of base pairs, then reassembling it in one fell swoop would succeed.1 On the brink of the project's beginning, skepticism remains. But an agreement between Celera Genomics Corp. of Rockville, Md., and the University of California at Berkeley's Drosophila Genome Project Group, may render that uncertainty irrelevant. The Berkeley group will essentially draft both a cheat sheet and a safety net for the Celera project, by completing the fruit fly's gene map and by generating about 12,500 unfinished Bacterial Artificial Chromosome (BAC) end sequences. The Celera group, in addition to producing raw shotgun data, will sequence the Berkeley BAC ends. Those two features will aid both teams in reassembling the Drosophila genome, notes Francis Collins, director of the National Human Genome Research Institute. "My sense is [reassembly] would not be possible" without the guide Berkeley's mapping and BAC end sequences will provide, Collins noted in a recent telephone interview. Gerry Rubin Venter, in another telephone interview, disagreed. "There's no question we could have done it without [the collaboration]," he says, but he adds that the joint effort will likely speed the project and ensure both its completeness and its accuracy--two components Collins has repeatedly doubted would arise from the shotgun sequence alone.2 Gerry Rubin, director of the Berkeley group, declined to speculate in a telephone interview on Celera's success had the company gone it alone: "I don't know. No one in the world knows," he comments. "This is not a matter for an opinion poll. This is a scientific experiment." However, he adds, combining the two techniques means that even if the Celera experiment by itself fails to reassemble Drosophila, the scientific community will still get its genome. "If the whole-genome shotgun approach works--great. If not, then we have this additional data." The agreement states that a successful public/private relationship on the fruit fly could also bode well for the Human Genome Project: "The parties believe their collaboration is more than a test of the scientific efficacy of the random shotgun approach to sequencing of the human genome. It is also a test of whether a public/private collaboration can expedite and lower the cost of the generation and use of genomic information ...," hinting that if the plan works, perhaps a similar agreement could be drafted for the human effort. "This is really a precedent-setting arrangement," Rubin says. Data Release Emerges, Again Venter first proposed the Drosophila project as a sort of test run for the large-scale shotgun approach that Celera will apply to the human genome. That plan drew criticism from around the world because the Celera human genome project would release data every three months. Participants in the international Human Genome Project require data release every 24 hours. Venter had called for public/private collaboration when he announced the Celera project.3 But Collins and others have said repeatedly that data release would be one obstacle preventing a closer working relationship. Still, Collins hinted that even without a formal agreement, the public and private data sets would benefit each other.4 The Berkeley-Celera agreement grapples with some of these data-release issues by calling for Celera to begin depositing contigs as unfinished sequences in GenBank upon completion of the random sequencing phase and by forbidding Berkeley from looking at Celera's data until it has been made available to all. However, GenBank users will likely not have the first look at Celera's raw Drosophila sequences, since it isn't required to make them immediately available. The company aims to generate income by selling subscriptions to its private database. Clients already include Amgen Inc. of Thousand Oaks, Calif. "I would guess Amgen wouldn't be willing to release data before they got a peek," Collins notes. "The early subscribers to the Celera database will get updates on all our data sets on a pretty consistent basis," confirms Venter, who adds that he offered the same consideration to Rubin and his colleagues. But NIH said it wouldn't fund Rubin if he had that favorable basis, because that wouldn't be consistent with the 24-hour release policy of the international Human Genome Project. "It's almost as if he's being punished because he wants to work with us and move things forward," Venter says. But Rubin acknowledges that having preferential access to Celera's data would hurt the international effort. "At first appearance, it may seem a little awkward," Rubin comments. "[But] it would be inappropriate for me to have the data when it would be inaccessible to others. Amgen Inc. www.amgen.com Celera Genomics Corp. www.celera.com Berkeley Drosophila Genome Project www.fruitfly.org/ The agreement, in effect, pushes Celera to deposit the data faster, because UC-Berkeley can't merge its data with the company's raw sequence information until then. "This is a reasonable compromise, which allows Celera to get a reasonable return on [its] investment," Rubin says, adding that Venter and Celera have been unfairly criticized for their information-release policy. Celera, unlike some other genomics companies, will at least make its data publicly available, albeit on a delayed and limited basis (the company also plans a Single Nucleotide Polymorphism database, which will likely be subscription-only). Critics "should save their criticism for other genomics companies [that] don't release anything." Perhaps one of the strengths of the plan, all parties agree, is that no one will have to change existing plans--much. "We're not doing different things than we were doing before. But we are doing them with different emphasis," Rubin notes. Changing emphasis on completing the Drosophila's gene map remains the most significant shift--and even that isn't much of a departure: "We were working on the physical map anyway," Rubin notes. And Celera, too, will not be significantly affected by the plan, since its Drosophila sequencing efforts haven't officially begun. The company's genome sequencing facility now has 30 sequencers online, with 18 more arriving shortly. Celera sequencers are calibrating those machines using small bits of Drosophila DNA. "I wouldn't characterize it as a full-scale assault on the Drosophila genome," Venter comments, hinting that sequencing could soon be under way without a formal announcement. Scientists at the company are also refining and testing its assembly software. "We can assemble the entire C. elegans genome in one day," he adds. Venter notes that once begun, the raw sequencing will proceed very quickly, with data available as early as June. Closing gaps and annotating the sequence--the last phase of the project, which will likely require the most active cooperation between both sides--will proceed slower. That end stage, Venter concludes, will prove two things: that publicly funded researchers and privately held companies can cooperate on genomics, and that mathematical models predicting the shotgun's success for Drosophila are correct. The first may be a necessity, because the second isn't a certainty, Rubin concludes. "You can do all the mathematical modeling you want, but [reality] has ways of not always obeying mathematics." P. Smaglik, "Privatizing the human genome?" The Scientist, 12[12]:1, June 8, 1998. P. Smaglik, "Private genome sequencing effort may hasten separate public venture," The Scientist, 12[14]:1, July 6, 1998. "Venter courts partners, denies data delays," The Scientist, 12[12]:6, June 8, 1998. P. Smaglik, "Human Genome Project deadline moves up two years, to 2003," The Scientist, 12[19]:1, Sept. 28, 1998

'Shotgun Wedding': Public, Private Drosophila Sequencing Agreement Should Speed Project, Ensure Accuracy

Interested in reading more?

Become a Member of