Bird Flu Fears Ignite Debate on Scientists' Sharing of Data

Although this cluster of cases appears to have the longest chain of transmission seen to date -- chicken to human to human to human -- preliminary analysis of the virus's genes has revealed no new mutations or evolutionary changes.

The RNA letters that make up the virus's genes are at the center of the dispute.

To a great degree, these "gene sequences" determine a microbe's behavior. Fewer than a dozen changes in flu's eight genes may be enough to give a strain pandemic potential -- the ability to spread easily between people who have no immunity to it.

An essential step in charting the spread of the H5N1 strain of bird flu in the past two years has been the analysis of gene sequences by a small number of labs in the United States, Europe and Asia.

Today, sequences from most newly collected samples of H5N1 are first deposited in the Influenza Sequence Database at Los Alamos National Laboratory in New Mexico. Access requires a password and is limited to about 15 research groups, mostly government and academic labs in WHO's global flu surveillance network.

Although everyone in this fraternity can use the data, the researchers who provide them get to publish the first scientific paper. The people who collected the virus or isolated it from a patient are usually listed as coauthors. Only after the paper appears is the gene sequence deposited in GenBank, a public database.

The details of the arrangement are murky, including the size of the Los Alamos repository and who ultimately controls access. Questions put to its manager, Catherine A. Macken, were answered with one sentence: "My policy . . . is to not reveal details about privately held data."

The alternative is to deposit sequences in GenBank as soon they come off the sequencing machine.

Run by the National Library of Medicine in Bethesda, GenBank is the world's largest collection of gene sequences (along with its European and Japanese counterparts, with which it shares data). About 2 million people around the world consult it each day. As of late last week, GenBank held 4,734 gene sequences from H5N1 influenza.

The issue gained public attention in February when Ilaria Capua, a 40-year-old virologist at the Tri-Veneto Region Experimental Animal Health Care Institute in Italy, sequenced the first H5N1 sample from Africa, isolated from a chicken farm in northern Nigeria. Someone at WHO invited her to contribute it to the Los Alamos data, but she declined and instead filed it in GenBank.

David J. Lipman, director of the National Center for Biotechnology Information, which oversees GenBank, said that since then, "we are seeing more movement toward people trying to cooperate and make data public. . . . Things are starting to open up."

Capua said this week that "what really surprises me is that I didn't say something that is completely out of context. I said, 'Wait a minute -- if this is the biggest threat, then we all have to run in the same direction.' "

Robert Webster, a leading flu virologist at St. Jude Children's Research Hospital in Memphis, doesn't disagree in principle. But, he said, "this is a difficult situation; there are many pros and cons to this one."

Most of the H5N1 samples his lab analyzes are provided by Asian scientists. Occasionally, they are given secretly if they come from a place where the virus has not been publicly reported. These collaborators have part ownership in the data, and Webster and others say they ask permission before loading the sequences into either the Los Alamos or GenBank databases.

Any system that immediately disseminates gene sequences must be one that Asian nations -- China in particular -- do not think exploits them.

Webster said he also feels a need to protect graduate students and postdoctoral fellows who do the sequencing work.

"I want to give them time to bring this information together [in a publication]," he said. "But if it's out there, the bottom feeders will use it, because there are bottom feeders."