Once the sequences were generated, Lipkin used computational techniques developed in his laboratory to filter out any remaining human sequences (which sometimes linger due to the presence of human RNA) and to piece together the many sequence fragments into longer strings. Of the more than 100,000 sequences initially produced, a mere 14 matched viral proteins in a database of all known microbes' sequences.

"If we had used a different sequencing strategy--one that gives you shorter reads--or if we had not used the sample preparation to enrich [for viral sequences], we would never have captured those," says Lipkin.

The virus from the patients' tissues was most closely related to a pathogen called lymphocytic choriomeningitis virus (LCMV), which is known to cause meningitis in humans. While LCMV has been implicated in transplant-associated illness before, the sequence of the new virus was different enough that existing methods could not have detected its presence. The results of the analysis were published online last week in the New England Journal of Medicine (NEJM).

Once it had characterized the LCMV-like virus, the group was able to design probes to test specifically for its presence. The group found evidence of the virus in several tissue samples from all three transplant recipients.

Unbiased high-throughput pyrosequencing has become a critical tool in Lipkin's lab, which is a member of the World Health Organization and helps train and equip public-health workers around the world. Lipkin has successfully used the technique to identify 20 viruses to date, including the Israel acute paralysis virus thought to be responsible for colony collapse disorder in bees. "There are all sorts of things that we've been able to identify using this approach," says Lipkin. "It's really quite powerful."

Because the sequencing technique is not biased toward known organisms, it is ideally poised to track down previously unknown pathogens. "We're finding the needle in the haystack, even without knowing what the needle looks like a priori," says Michael Egholm, vice president of research and development at 454 Life Sciences and a coauthor of the NEJM report.

"There's an enormous amount of uncharted territory in microbiology," says Lipkin. As many as 40 percent of cases of central nervous system disease cannot be traced back to a specific culprit. For respiratory illness, the figure is 30 to 60 percent. In the United States alone, 5,000 deaths each year result from unidentified food-borne infections. "The advent of molecular tools like the one we've described here will be important in identifying the pathogenesis of a wide variety of diseases, acute and chronic," says Lipkin.

According to Whitley, understanding the microorganisms that cause these diseases could lead to more effective treatments.

As powerful as 454 sequencing is for discovering new pathogens, it is not fast or cost efficient enough for use in routine screening of transplant tissue. But microbes discovered using this technique could be incorporated into existing screening techniques. "As we do more and more transplantation medicine," says Lipkin, "it's going to become critical that we find faster, more efficient, less expensive ways to screen to ensure safety."