But other researchers wonder if the consortium is breaking up prematurely. They believe a historic opportunity has been lost to build and maintain a distinctively collaborative approach to genomics.

Most of the human genome was sequenced at four centres in the United States and one in the United Kingdom. But many other universities and centres in America, Europe and Japan also took part. The links between the partners have been breaking up more rapidly than some of them had hoped.

Ironically, one of the things driving the sequencers apart is the whole-genome shotgun sequencing technique that once brought them together. The announcement in 1998 by Celera Genomics, based in Rockville, Maryland, that it would use the technique to sequence the human genome before the HGP, helped to galvanize the publicly funded consortium.

Whole-genome shotgun sequencing lends itself to a centralized approach, as any facility with enough sequencers can break the genome into millions of pieces, read them and reassemble them. In contrast, the HGP divided the genome into small segments, assembling the pieces soon after they were read.

The US National Institutes of Health (NIH) tacitly acknowledged the merits of Celera's approach when it revised its plans to sequence the mouse. Initially, the NIH considered applying the same clone-by-clone technique used by the HGP. But the National Human Genome Research Institute has decided instead to sequence most of the mouse genome by the shotgun method, using the clone-by-clone approach only to build a map to help piece the shotgun data together.

 

SAM OGDEN

Gene shift: Lander says partners will compare notes, not share projects.

The validation of Celera's technique — and the build-up in the number of sequencing machines at each major centre to prevent Celera from taking the lead — means that individual centres will soon be able to decipher small genomes on their own. The UK Wellcome Trust's Sanger Centre, near Cambridge, is taking on the zebrafish genome alone (see Nature 408, 503; 2000), and the US Department of Energy's Joint Genome Institute in California sequenced 15 bacterial genomes in an October "Microbial Marathon" that also served, in effect, as a declaration of independence.

Richard Gibbs, director of the sequencing centre at Baylor College of Medicine, Houston, says that the loosening of the consortium was inevitable. But he thinks this is happening too quickly. "It makes most sense to enjoy the fruits of its efficiency as long as possible," he says.

Bob Waterston, however, director of the sequencing centre at Washington University in St Louis, says that working alone makes projects easier to manage.

Eric Lander, director of the Whitehead Institute at the Massachusetts Institute of Technology, does not foresee the consortium disbanding completely. Instead, he suggests it will move from sharing projects to discussing the merits of different sequencing strategies and technologies.