The genome duel seemed to have subsided until one side took a vicious stab at the other this week. In a paper published in the online 5 March issue of the Proceedings of the National Academy of Sciences, three leaders of the publicly funded, international Human Genome Project (HGP) assert that what appeared to be a dead-heat race to sequence the genome was actually nothing of the sort. Rockville, Maryland-based Celera Genomics, the authors argue, did little more than gather long tracts of sequence from the HGP and break them down into patterns that were easily reassembled.
In the race to sequence a draft of the human genome, Celera used a so-called whole-genome shotgun approach. It chopped the entire genome into tiny pieces and then reassembled them all at once. In contrast, the international consortium adopted a more incremental strategy, piecing together smaller amounts of data. Both sequences were published in February, 2001; Celera's in Science, and the HGP's in Nature.
The three PNAS authors--Robert Waterston at Washington University in St. Louis, Missouri; Eric Lander at the Massachusetts Institute of Technology in Cambridge; and John Sulston at the Wellcome Trust Sanger Centre in Cambridge, United Kingdom--sought to mimic Celera's breakdown and reconstruction of the HGP data--which the company has always admitted combining with its own data.
The trio simulated three shreddings of a chromosome, two of which were random and one of which resembled Celera's disassembly pattern. The latter, they found, yielded on reassembly a sequence essentially identical to the original; the other two gave a fairly nonsensical jumble of DNA. With Celera's approach, they conclude, you get out what you put in--in this case the HGP data. Phil Green of the University of Washington in Seattle, concurs. "It is important to correct the historical record," he says.
"They say that we copied their answer, and that's completely false," says Mark Adams, vice president of genome programs at Celera. The two teams' versions of the human genome sequence were different, he notes, and he says that Celera kept far less genomic data intact than the PNAS paper suggests.
Others question the need for rehashing old rivalries. "This [paper] shouldn't be taken to represent the views of the entire public effort," says Richard Gibbs of Baylor College of Medicine in Houston. Gibbs took part in the HGP and is now collaborating with Celera on the rat genome. "You get away from the recognized process of science when you start this arbitrary deconstruction of other people's work," he says. Consensus, apparently, is a long way off.