Britain's Wellcome Trust and several companies are discussing the creation of an international consortium to advance structural genomics — the determination of the structures of proteins, RNA and other biological macromolecules.

Currently, most structures are deduced slowly and expensively by researchers studying a particular macromolecule. With the draft sequence of the human genome now available, the consortium's eventual goal would be to complement this approach with an industrial-scale drive to develop the high-throughput determination of thousands of structures.

Williamson: champion of structural consortium.

Alan Williamson, former vice-president for worldwide research strategy at Merck and now an independent consultant, is championing the project. An interim committee has been created to study its feasibility.

Williamson played an important role in brokering the so-called SNP Consortium, a non-profit foundation that aims to put 300,000 single-nucleotide polymorphisms (SNPs) — differences in single DNA base pairs between individuals — into the public domain by the end of of 2001.

The companies on the interim committee have not been named. But some of the 13 who are members of the SNP consortium — including IBM, one of the latest recruits — are believed to be involved.

“An effort to significantly increase the number of target protein structures in the public domain would be of great interest to anyone involved in rational drug design,” says Rob Cook, head of molecular recognition at one of these companies, GlaxoWellcome in Stevenage, England.

Shape of things to come? An industrial-scale attack on protein structures — such as the growth hormone somatotropin, shown here — could aid drug design. Credit: ALFRED PASIEKA/SPL

“I was approached by several companies interested in increasing the rate [at which] structures flow into public-domain databases,” says Williamson. “Now we want to open up the consortium to others.” A decision on whether to go ahead might be taken by the end of the year, he says.

The cost of consortium membership has not been fixed, but is likely to be around US$1.5 million per year. Structures would be determined at public research institutes.

“We certainly see the consortium as potentially a very good way to further structural genomics,” says Michael Morgan, chief executive of the Wellcome Trust's Genome Campus near Cambridge, England. “We have an ‘in principle’ agreement with the governors of the trust that if we can get the consortium to work, [the fund] will chip in to the level needed to make it a success.”

The move reflects growing academic interest in taking structural genomics to an industrial scale, although views differ on when and how this should be done.

At a meeting in Cambridge in April, organized by the US National Institute of General Medical Sciences (NIGMS) and the Wellcome Trust, scientists from public agencies in nine countries agreed to form an ‘International Structural Genomics Initiative’.

This initiative aims to coordinate national activities, avoid duplication of effort, and adopt common data-release policies. Its goal is to produce 10,000 protein structures in ten years.

But the initiative does not dictate targets and strategy or fund structural centres. It may evolve to do this, says John Norvell, head of the US Protein Structure Initiative. But before this can happen, many technical and logistical issues need to be resolved.

The Protein Structure Initiative hopes to deal with these issues; it will announce its first round of grants to a handful of pilot centres in the United States next month.

The three existing large programmes in structural genomics are run by the NIGMS, the Institute of Physical and Chemical Research (RIKEN) in Japan, and the Wellcome Trust.

Norvell agrees that a consortium could speed things up. “The participation of industry in the international effort is strongly encouraged,” he says. But the NIGMS has not joined the consortium. “It's early in the game for all of us,” says Norvell.

The Wellcome Trust's participation in any consortium is conditional on “data being made freely available”, says Morgan. He adds that all the companies in the consortium are against intellectual property being taken on any data that it produces.