Some say it's a nightmare scenario out of the old movie Outbreak with Dustin Hoffman or the newer Contagion by Steven Soderberg: a pandemic caused by the accidental release of a man-made strain of lethal H5N1 bird flu virus.
Could such a disaster happen in real life?
The question is serious and timely. Two teams of researchers who have created such a deadly viral strain agreed last month to halt their work for 60 days, because of concerns that bioterrorists could turn the engineered virus into a weapon.
The two research teams, in the Netherlands and the United States, announced in September that they had independently created an H5N1 strain that was transmissible between people and lethal in more than 50 per cent of cases.
Their virus has been transmitted between ferrets that merely breathed the same air, which is a clear indicator to virologists that the strain could travel easily between people.
Such a flu virus may spread as rapidly as ordinary seasonal flu, but it is several degrees of magnitude more deadly.
The H5N1 strain of bird flu is fatal in 60 per cent of human cases.
One of the studies, led by Ron Fouchier of the Erasmus Medical Centre in the Netherlands, has been submitted to the journal Science. The other, led by Yoshihiro Kawaoka of the University of Wisconsin, has been sent to Nature.
The US National Science Advisory Board for Biosecurity made an unusual request last month, asking both research teams not to publish details about how they created the more virulent strain of H5N1, out of concern for security.
The World Health Organisation will hold an international discussion about the best way to proceed, starting on February 16.
Many questions need to be discussed: do the laboratories involved have enough biosafety protection to make sure the killer strain cannot escape?
And does the value of such work outweigh the risks? The scientists defend their research as an effort to better understand the potential for H5N1 bird flu to mutate into a form that can spread easily between people through coughing or sneezing. Indeed, finding the biological determinants that make a virus transmissible between humans is one of the most important scientific goals in the field.
Defining such markers may help us to distinguish particularly dangerous emerging viruses, from the huge diversity of animal viruses that pose little or no threat to humans. The studies should also improve surveillance efforts in detecting similar changes in natural strains, and to develop diagnostics, drugs and vaccines.
Both experiments were conducted in labs with security at the biosafety 3 (BSL3) level. Scientists must shower and change their clothes when leaving the labs, whose safety features include negative air pressure and high-efficiency particulate air filters to clean the exhaust air. That should be enough protection to ensure against an accidental release of the virus.
Some ask if such work should be done in labs with the more stringent BSL4 protocol. But that would hamper the research needed to develop countermeasures against an H5N1 pandemic, says Masato Tashiro, a virologist at the National Institute of Infectious Diseases in Tokyo - because it would limit the number of researchers able to work with the virus.
Others argue that the new H5N1 variants must be restricted to BSL4 labs, to protect not only the researchers working on the viruses, but also society at large. Many of us in Hong Kong remember that severe acute respiratory syndrome (Sars) accidentally infected staff at four high-containment labs in mainland China, Taiwan and Singapore, which were BSL3 and BSL4 facilities.
One big difference from Sars is that there are H5N1 vaccines that the lab workers can be given - to protect them from infection and reduce the chance of a virus escaping from the lab.
Fouchier and his colleague Ab Osterhaus, the Dutch flu researchers, say their work will make it easier to scan animal populations for similar mutations. This will provide an early warning about a potential outbreak and help specialists prepare for a possible pandemic.
A vital area of work is understanding the mechanisms that allow an animal virus to be transmitted to humans, and the scientists argue their mutations represent clearly identifiable biological traits.
The article sent to Nature says: 'Taken together, we strongly feel that the benefits of sharing this state-of-the-art data by far outweigh the potential risks ... The primary goal of our research has direct short-term and mid-term implications for awareness of the pandemic potential of H5N1 viruses, surveillance and pandemic preparedness.'
Ordinary people ask: Are we safe? Should these experiments be stopped immediately?
One international authority who defends the research is Professor Malik Peiris, a virologist at the University of Hong Kong. It is vitally important, he argues, to understand the biological factors that allow an animal virus to spread to and between humans.
'Biosecurity is often equated with BSL3 and BSL4 and space suits,' Peiris says. 'Far from it. Our grandfathers worked with smallpox, polio, Japanese encephalitis, dengue on the open bench - what we would now call BSL1 or BSL0 - and 99.99 per cent of the time, they did so safely.
'This is not to decry the need for biosafety. Now that we have the technology, we would like to prevent that 0.01 per cent when things go wrong. But at the end of the day, it is the training and not so much the hardware - the BSL grade - that counts.'
The wisdom of the international community will find a proper balance between security and understanding, he says.
But there is no doubt the danger must be taken seriously: the research should be transparent not only to other researchers, but also to the public.
Could terrorists or rogue states use the published virus data to cause havoc? Peiris believes small terrorist groups would not have the technology to genetically engineer such a virus. A state that wanted to do so could probably produce a virus without the help of the information in the two magazine articles, he said.
Past experience, Peiris says, reminds us that nature is more efficient at creating dangerous viruses that we humans could ever be. It is an illusion, he believes, to halt such research out of a concern for safety.
So we face one of those dilemmas that dot the history of science and technology. The wise Winston Churchill characterised such situations very well: 'Humanity has gone too far forward to return, and moves too quickly to stop.'
Viruses, however, move even faster than we do.
Reinhard Renneberg has been a professor of bioanalytical chemistry at Hong Kong University of Science and Technology since 1994