Paris

Celera Genomics, the private-sector rival to the publicly funded international effort to sequence the human genome, is now said to be contemplating an industrial-scale Human Proteome Project, this time well ahead of any comparable public venture. The goal would be to identify the properties and functions of every protein expressed in an organism: its ‘proteome’.

Celera is in advanced discussions with Denis Hochstrasser, one of the founders of GeneBio — the commercial arm of the Swiss Institute of Bioinformatics, which is among the world's leading protein research centres — and an authority on two-dimensional polyacrylamide gel electrophoresis (2D PAGE) technology. The goal is to combine new technologies with the workhorse 2-D electrophoretic gel technique to develop an industrial-scale attack on the proteome.

Reaching a deal now seems to hang on the question of data access. Amos Bairoch, a co-founder of GeneBio, says that joining up with a private company will result in restrictions on data being placed in the public domain. He would prefer the Swiss groups to collaborate with a publicly funded international human proteome project, with all data being put rapidly into the public domain.

As the sequencing of the human genome nears completion, characterizing the human ‘proteome’ is the next big task (see Nature 402, 715–720; 1999). One approach is to try to predict proteins directly from gene sequences. In practice, however, this is riddled with problems.

Post-translational modifications, for example, mean a single gene sequence can yield over 20 structurally different proteins. So researchers instead want to identify the actual proteins found in any biological sample.

The many difficulties in scaling up proteomics, however, have made an industrial-scale approach impractical up until now. Protein spots are often still excised by hand from gels, for example, for further analysis, while hydrophobic or big proteins do not dissolve in the solvents used in 2-D PAGE.

That may now be about to change. Last November, Hochstrasser and colleagues at the Swiss Institute of Bioinformatics published a blueprint for a ‘molecular scanner’ that simultaneously digests proteolytically the thousands of proteins on a gel, and electro-transfers them onto a membrane, in a procedure akin to Southern blotting (see Anal. Chem. 71, 21; 1999).

The peptides on the membranes are then measured by laser mass spectrometry, and the protein identified by matching this ‘peptide mass fingerprint’ automatically over the Internet with entries in a protein sequence database. The result is an annotated digital map of the proteins on a gel — a snapshot of any sample — that is available online. This database can be searched by protein name, or by peptide fingerprint.

Hochstrasser says Perkin-Elmer Corporation, which produces mass spectrometers, has agreed with GeneBio to develop the ‘molecular scanner’. Significantly, Perkin-Elmer is also about to release a next-generation series of mass spectrometers, touted as being a major improvement on existing machines.

Meanwhile, Craig Venter, president of Celera Genomics — a subsidiary of Perkin-Elmer — has been wooing Hochstrasser with a view to a joint venture. This may combine hundreds of the new machines with the Swiss technology in a factory approach that would seem to amount to nothing less than an all-out human proteome project.

“Nothing has been signed yet,” says Hochstrasser, adding that GeneBio, which is committed to public access, would not be involved. A “new entity” would channel the Swiss groups' expertise, he says. But, he adds, “There is no question [that a deal with Celera] would lead to a major scale-up.”

He reckons that the problem of hydrophobic proteins can also be overcome by first precipitating them and redissolving them in alternative solvents. But, he says, the complexity of proteomics means that, although an industrial-scale approach could yield enormous amounts of data, alternative technologies will also be needed to finish the job.

Bairoch would prefer the Swiss institute to work with a publicly funded international consortium along the lines of the international human genome project. “It is a big debate in Geneva,” he says. But he admits that this may be difficult. Whereas the US National Institutes of Health is playing the largest role in the human genome effort, Europe is strong in proteomics and would probably demand a leading role in any international effort.

Bairoch: fears loss of public access.

At the same time, the political fragmentation of Europe means that there is little realistic hope of Europe getting organized quickly, says Bairoch. He points, for example, to the fiasco when the European Union almost let its foremost bioinformatics structure, the European Bioinformatics Institute, fall into bank ruptcy (see Nature 402, 3–4; 1999). And when the publicly funded Swiss-Prot, one of the world's leading protein databases, faced closure for lack of money (see Nature 381, 266; 1996), it was only rescued by the creation of GeneBio, which ploughs 75 per cent of its profits back into the Swiss Institute of Bioinformatics.

A deal between Celera and Hochstrasser might prod the public sector into action, he adds. “I think the combination of having Celera in the human genome race was ultimately beneficial for all,” says Robin Offord, a director of GeneBio.

Hochstrasser says, “We will try to make as much available for free as we can.” But he argues that a compromise — such as an agreed delay in releasing data — may be needed to satisfy the demands of the private sector. “It is difficult to find the balance,” he says. “We haven't signed yet because we would like free access to protein data. That is why it won't be GeneBio, but a new outfit.”

Officials from the human genome project in the United States and Britain said comment would be premature since this was the first they had heard of a private-sector human proteome initiative.

Neither Celera nor Perkin-Elmer was available for comment. Heather Kowalski, corporate communications manager at Celera Genomics, explains that the companies are in a “quiet period” in the run-up to a secondary stock offering of 1.6 million Celera shares in early March.

The prospectus states that the offer is to fund “Celera's new product and technology development activities in functional genomics, with an emphasis on proteomics,” as well as “general corporate purposes including possible acquisitions, alliances or collaborations”.