Geneticists and many others have cause to celebrate this week. The consortium behind the publicly funded international Human Genome Project has just put the one-billionth base pair of sequence into the public databases, and today is publishing the first sequence of a human chromosome, number 22 (see pages 447, 467 and 489). Given the scale and complexity of the task, the sequence is a great achievement.

The consortium's major rival, Celera Genomics, has itself also just completed the 180 million base-pair genome of Drosophila melanogaster (see Nature 401, 729; 1999). It did this in collaboration with an academic team headed by Gerry Rubin at the University of California at Berkeley, funded by the US National Institutes of Health (NIH). Participants in the collaboration saw it as a precursor for similar cooperation on the human genome. Their agreement stated that it was “also a test of whether a public/private collaboration can expedite and lower the cost of the generation and use of genomic information”.

“Yes” seems to be the answer. Rubin's group provided Celera with maps of the fly genome, for example, and Celera quickly produced additional high-quality sequence data. It has been a win–win affair.

Tensions have so far blocked public/private collaboration on the human genome. Talks between Celera and the US Department of Energy foundered late last year on the thorny question of data access (see Nature 397, 93; 1999). The NIH and Britain's Wellcome Trust opposed the deal as conflicting with the so-called “Bermuda agreement” under which sequence data obtained through the Human Genome Project are released into public databases within 24 hours (see http://www.sanger.ac.uk/HGP/policy-forum.shtml#ref1, which includes a concisely readable description of the sequencing process).

Craig Venter, Celera's president, has promised to release human sequence data quarterly — his subscribers have immediate access. The company argues that the Bermuda agreement is less relevant than it was. What Celera is seeking to patent — as indeed are many scientists at NIH and elsewhere — are genes of demonstrable utility, rather than raw sequence data and genes of unknown function, whose patenting the Bermuda agreement was indirectly designed to pre-empt.

Rubin and Celera managed to strike a pragmatic compromise that included Celera depositing data in GenBank on completion of sequencing. But the value of the cooperation rested less on simply seeing data, and more on the fact that scientists from both sides worked together on the strategy for assembling the sequence data and identifying the genes they contain. The result has been faster and better for it.

It is increasingly clear that the different approaches being taken by Celera and the Human Genome Project are complementary. Collaboration could provide increased coverage and accuracy of the human genome and a higher rate of discoveries important for human health; importantly, a comparison of the two sequences would also yield many single-nucleotide polymorphisms, the major form of DNA variation involved in human traits and diseases.

More substantial collaboration would certainly accelerate completion of the human genome. One question is: at what price in future licensing requirements that would impede research? No progress has been achieved in bridging the public and private genome projects, mainly because the current policies on data release are still far apart. Venter's commitments to his investors are his affair; if he and the Human Genome Project participants can find a mutually acceptable moral and expedient policy on data release, the significant benefits of closer cooperation between Celera and the publicly funded project would be realized. That possibility should be pursued.