Washington

In a bid to crack protein structures, the US National Institute of General Medical Sciences (NIGMS) last week launched the largest project to date to find the best methods for unravelling the molecules' three-dimensional make-up.

As part of its structural genomics initiative, the institute will provide $150 million over five years to seven regionally based research groups. The NIGMS aims to solve the three-dimensional structure of 10,000 proteins — each representing a protein family — over the next decade.

Structural genomics compares how proteins of known structure share functions or gene sequences with unsolved proteins. It then uses computation to predict structures for the unsolved proteins.

But to assemble a full protein library, structural biology's technology and methodology need to improve. To that end, the seven groups will use robotics to automate protein characterization. Some will also seek to expedite or eliminate steps in sample preparation — the biggest bottleneck in large-scale characterization.

Each of the groups is using a different rationale for choosing which proteins to characterize. They will study proteins from different organisms, ranging from yeast to humans. Some will use X-ray crystallography, whereas others will rely on nuclear magnetic resonance spectroscopy to solve the structures. The challenge will be to take advantage of the diversity of approaches while avoiding redundancy, says John Norvell, director of the NIGMS project.

But some duplication of effort is desirable, as similar structures can arise from drastically different genetic sequences. It would also be helpful when groups are trying out new methods for discerning structures.

In addition, redundancy could come in the form of competition from similar projects. For example, Britain's Wellcome Trust is thinking about organizing an industrial structural genomics consortium (see Nature 406, 923; 2000).

But Stephen Burley, a Rockefeller University structural biologist, and one of the seven principal investigators selected by the NIGMS project, says that industrial researchers will be more interested in medically important proteins. The NIGMS project, in contrast, aims for breadth, not depth.

⇒ http://www.nigms.nih.gov/funding/psi.html