Little Consensus at Scientific Meeting on Releasing Results of Risky Research
A two-day meeting here brought scientists no closer to resolving the question of whether there are any kinds of experiments whose results should be kept from the public.
Arriving at an international consensus about whether scientific journals should occasionally publish censored versions of papers because the full ones might prove useful to terrorists “is likely to take multiple years,” Bruce Alberts, the editor of Science, told an audience at the Royal Society.
The meeting on Tuesday and Wednesday was called after Science and its English rival, Nature, had each agreed late last year to hold off publishing papers on lab-engineered versions of the bird flu virus at the request of the U.S. government. The journals last week got the go-ahead to print the papers.
“My fear is that now this crisis is over, nobody will work on this” problem of whether to publish the results of similar experiments in the future, Alberts said.
The issue touches on science’s culture of openness, the funding of research, global equity, cybersecurity and the effect of regulation on human behavior.
The papers originally in question described successful efforts to enhance the transmissibility of the H5N1 avian influenza virus, a pathogen that rarely infects people but when it does kills them about half the time. One experiment was conducted by Ron Fouchier of Erasmus Medical Center in Rotterdam. The other was done by Yoshihiro Kawaoka of the University of Wisconsin and the University of Tokyo.
The request not to publish came from the National Science Advisory Board for Biosecurity, a committee of scientists that advises the U.S. government about federally funded research, which both studies were. The panel changed its mind last Friday after learning more about the specifics and the relevance of the experiments.
The audience of about 200 scientists and ethicists considered numerous questions.
Are there experiments so pointless and dangerous that they should never be done? Are there experiments that might be useful but are too dangerous to be done? Is there any purpose in reporting that something has been accomplished without saying how it was done?
There was general agreement that some experiments are off limits, such as attempting to make the AIDS virus transmissible by air. There was less agreement about the experiments at hand, which changed the characteristics of H5N1 bird flu.
A common technique for learning what a gene does is to knock it out, which usually results in a less dangerous microbe. Fouchier said that strategy is inadequate for trying to discover small genetic changes that alter a flu virus’s behavior.
“Why does a car drive?,” he asked, making an analogy. “You can take out many parts and the car will not work. That doesn’t tell you much.”
He and Kawaoka, working separately, used the opposite strategy. They either engineered in or indirectly encouraged mutations that enhanced the contagiousness of H5N1 influenza.
“The only way of increasing our knowledge is by this ‘gain-of-function’ approach,” Fouchier said.
The rationale for the work is that it will give virologists and epidemiologists an idea of what worrisome mutations to look for when they analyze samples of H5N1 viruses collected from birds around the world.