A new initiative on sharing avian influenza data has just been announced, called the Global Initiative on Sharing Avian Influenza Data (GISAID). This is the latest in a series of developments that have opened up influenza sequence data to the world scientific community to an unprecedented extent, a dawning recognition that the usual rules of scientific behavior and etiquette about who “owns” data need modification in the face of a potential devastating pandemic, where weeks or months head start on a scientific pathway can make an appreciable difference.

Last month Indonesia gave permission for sequencing laboratories in Hong Kong and at CDC in the US to release data collected from its isolates. Two days ago US CDC released some 650 US influenza sequences (human subtypes from seasonal influenza) it had been sitting on for a considerable periods. Some veterinary virologists had earlier made a public commitment to release their data. So GISAID is not by any means in the forefront, but it may be one of the most important of the initiatives, at least with respect to human H5N1 sequences.

As far as we understand it, on the basis of a quick read and without the kind of clarification that will be forthcoming in the next few months as it is put into operation, scientists who join the collaborative will be required to place their sequences in a secure database as soon as possible after producing and validating them. Collaborating scientists will thus see everyone else’s data immediately as it is produced, but in return, agree to publish collaboratively or jointly. As soon as the data is analyzed it is to be deposited in one of the three publicly available databases participating in the International Sequence Database Collaboration (EMBL, DDBJ and GenBank) — irrespective of whether it had been the subject of a publication, and in no case later than 6 months after submission. The qualifications for admission are not yet clear. Is a bioinformatics analyst who doesn’t produce data barred from entry, for example, eligible for membership in the collaboration? GISAID will be governed by an international panel of distinguished scientists, who presumably will be gatekeepers, a possible conflict of interest. To the extent their decisions are transparent, we believe this should not be a major problem. To the extent their decisions are seen as arbitrary or are secret, this is a possible weakness in the current plan. We shall see.

Scientists who have been accepted into GISAID will share their sequence data immediately in exchange for authorship credit when their data are used; and they will be obligated to allow their data to be publicly available within 6 months of submission to the GISAID secure database. The framers of the new agreement believe this lag time will become shorter as experience and confidence with the system accrue.

This is clearly an improvement over the current situation. As we read it, it would require immediate access to many of the sequences currently in the LANL secure database, or at least the ones older than six months, and the publication of all of them within six months and the publication of those done by collaborating scientists from now on. It would also seem to allow many more scientists to become collaborators than are now part of the notorious Gang of Fifteen who currently guard the LANL secure database. Opening up the sequences to a wider range of scientists is an important goal, and this will happen much more quickly under GISAID. Those not part of the collaboration will see the sequences no later than six months after they are produced by the laboratory. We would prefer a shorter interval, or even immediate public access, but we’ll take what we can get at this point. We are glad to see substantial progress and we’ll see how this works.

This doesn’t seem to address the concern of many developing countries that information from isolates obtained from their people or animals will be used for commercial purposes (e.g., a vaccine) that will then be too costly for them to afford. One way this could be solved would be to require any vaccine produced from a seed strain from a country like Indonesia or Vietnam to be provided at cost or less to the originating country. This would also encourage countries to release sequence data.

All in all, this has been a good week for the ideal of scientific openness. We give a tip of the hat to those who have worked hard to forge an agreement on a difficult issue. While those outside the area may feel the intellectual property, authorship and priority issues of scientists and countries are minor concerns compared to the threat of a pandemic, those issues exist and are significant barriers to the sharing of all kinds of scientific data. They are not easily solved problems and we applaud everyone involved in the GISAID who rolled up their sleeves and set to work to solve some of them. This is a first try and we expect there will be further progress.