Research into the H5N1 virus has to be carried out in highly controlled conditionsExperts have delayed a decision on whether controversial research into the H5N1 bird flu virus should be released.
It had been looking at how the work could be released while guarding against its abuse by bioterrorists.
But talks at the World Health Organization in Geneva decided more discussions were needed to see if it could be possible to publish in full.
One of the two journals which want to publish has already agreed to wait for talks to be complete.
Mutation fearsThe controversy is centred on two research papers - one of which was submitted to Science, the other to another leading journal, Nature, last year.
They showed that the H5N1 virus could relatively easily mutate into a form that could spread rapidly among the human population.
The studies prompted the US National Security Advisory Board for Biotechnology (NSABB) to ask both journals last November to redact some sensitive parts of the research, which it believed could be used by terrorists to develop such a virus.
Continue reading the main storyAnalysis
David ShukmanScience editor, BBC NewsThese talks go to the heart of a fundamental debate over whether scientists should operate openly and publish all their findings - which is a basic principle of modern research - or whether some subjects are so sensitive that some key details should only be available to a carefully vetted audience.
The talks in Geneva have not resolved this.
The researchers passionately believe that the best way to tackle the threat of a pandemic is to understand how the virus can mutate, and that only by releasing their results in scientific journals will progress be made.
Ranged against them are experts in security who argue that too much information in the public domain will create another weapon for terrorists.
It's a highly sensitive dispute - the scientists fear that any kind of censorship will set a precedent of government control over their work.
So it's no surprise that another round of talks will be held in a couple of months, and in the meantime, research will remain on hold.
The request caused outcry among some scientists who believed that it was an infringement of academic freedom.
Some pointed out that the scientists had given presentations about their work at conferences and the details were already widely circulated, so redaction would have little purpose.
The scientists who carried out the research, and the journals concerned, have been considering the request and listening to suggestions as to how the research results could be redacted in the scientific journals, but distributed to bona fide researchers who urgently need the information.
The information is vital to develop a vaccine to any human form of bird flu, and it would enable surveillance teams to see if the bird flu virus was mutating into a form that could be transmissible to humans.
But such efforts have been put on hold for four months as governments, scientists and the journals decide what to do.
The Geneva meeting of 22 scientists and journal representatives agreed that publishing only parts of the research would not be helpful, because they would not give the full context of a complete paper.
It agreed to extend a temporary moratorium on research using lab-modified H5N1 viruses, but also recognised that research on naturally occurring virus "must continue".
'Critical importance'Dr Keji Fukada, assistant director-general of health security and environment for the WHO, said: "Given the high death rate associated with this virus - 60% of all humans who have been infected have died - all participants at the meeting emphasised the high level of concern with this flu virus in the scientific community and the need to understand it better with additional research.
"The results of this new research have made it clear that H5N1 viruses have the potential to transmit more easily between people underscoring the critical importance for continued surveillance and research with this virus."
Dr Fukada added: "There is a preference from a public health perspective for full disclosure of the information in these two studies. However there are significant public concern surrounding this research that should first be addressed."
Experts will now look at what information is already in the public domain and how that relates to the contents of these research papers.
A further meeting is likely to happen in a couple of months' time.
Nature has said it is happy to wait - if there is a chance it will able to publish in full.
Science's editor Dr Bruce Alberts, had previously said it also wanted to publish full details of the work, unless progress was made on how to circulate details of the findings to scientists.
Doctors want to improve the prognosis of this aggressive cancer
