Transplanted organs can carry viruses. Credit: AJ PHOTO/SPL

The discovery of a virus that may have killed three transplant recipients in Australia could mark a dramatic acceleration in the speed at which new pathogens can be identified. But it raises concerns that the ease with which such suspects can now be found could lead to researchers overlooking the need to firmly establish them as the cause of the disease in question. Fingering the wrong microbe could lead to inappropriate treatment or divert attention away from the real cause.

The three patients received organs from a single donor in Melbourne in December 2006. By January all three were dead. Ian Lipkin of Columbia University's Mailman School of Public Health in New York, who specializes in identifying new pathogens, heard about the case and collaborated with the Australian researchers who first took it on. After looking for thousands of telltale signs of known pathogens without success, Lipkin, who sits on the advisory board of 454 Life Sciences in Branford, Connecticut, decided to use the company's technology to sequence genes from samples that had been filtered to enrich their non-human DNA component.

Roughly one month and 144,000 fragments of sequence later, a homemade algorithm in Lipkin's lab had pulled out 14 gene fragments that looked viral. Preliminary, unpublished analyses suggest they come from a new member of the family Arenaviridae. Although this sort of sequencing has been used to identify viruses in the past, the 454 technology cuts down on time and effort, says Anthony Fauci, director of the US National Institute of Allergy and Infectious Diseases, which funded part of Lipkin's work. Lipkin estimates that the technique could be used to process samples in a matter of days.

But simply finding a virus is not enough, cautions microbiologist David Wang of Washington University in St Louis. Establishing that the virus actually caused the deaths is also critical, and is a lot harder. At present, the virus has been found in tissue from all three patients, and not in tissue from 60 controls. The virus's closest relative seems to be the lymphocytic choriomeningitis virus, which is believed to have killed organ-transplant patients in the past. But the team does not know whether the virus was also present in the donor. Lipkin says donor tissue samples have not been supplied to him by his Australian collaborators.

The guilt-by-association approach to pathogens can be misleading, warns microbiologist David Relman of Stanford University. Viruses do not necessarily behave in the same way as their closest known relatives. And the very fact that these new techniques work with raw sequence, rather than entities that could be grown and studied in the lab, makes follow-up experiments more challenging.

“It's not hard to find somebody you can implicate,” says Relman, in police-procedural mode. “What's really hard is to nail the conviction.”