One of the question marks hanging over cooperation in structural genomics is how and when the structures of proteins and other macromolecules should be made public.

At a meeting in Cambridge in April, organized by the Wellcome Trust, scientists from public agencies in nine countries agreed that there should be timely release of all data.

But a loophole remains in that, for the next couple of years, the decision on when a structure is complete and ready for a database is left to researchers themselves. This is because there is no agreement or way of defining when a protein structure is accurate and complete — it is largely a matter of personal judgement.

Arriving at an automatic cut-off point based on “numerical criteria” of quality is “now a key goal”, says John Norvell, head of the US Protein Structure Initiative based at the National Institute of General Medical Sciences. This would allow for clear rules on the timing of data release.

An early agreement between researchers on data access is also critical to facilitating collaboration with industry. As several delegates at a meeting organized by the Organisation for Economic Cooperation and Development in Florence in June pointed out, researchers may be reluctant to put valuable structural data into the public domain.

The Cambridge meeting agreed that data released should be accompanied by a short research paper, published simultaneously on the Internet.