UF biologists join world experts in experiment to explore flu origins

May 6, 2009

GAINESVILLE, Fla. — As scientists rush to figure out how the virus responsible for the current influenza outbreak originated, a group of the world’s experts on evolutionary biology — including a University of Florida team — has set aside some academic traditions to work together toward a solution.

Rather than rush to publication with their individual findings on the current flu outbreak, the computational biologists have opened their work in progress to each other and the public through a Web site: http://tree.bio.ed.ac.uk/groups/influenza.

Researchers in Britain came up with the idea to rapidly release data analysis results. The site allows a broad collaborative yet independent effort by allowing registered users to post their results and comment on the work of others. Contributors hail from institutions around the world, including the University of Oxford, the University of Edinburgh, the University of Florida, the University of Arizona, the University of California-Los Angeles, the University of Hong Kong and Belgium’s Rega Institute.

“It’s a really new concept, because basically there is a worldwide emergency, so this is a very fast way to have data circulating very quickly in the scientific community,” said Marco Salemi, a UF assistant professor of pathology immunology and laboratory medicine, who has contributed results to the Web site.

Site creators caution that analyses are preliminary and corrections likely will be needed.

Still, there is great value and a certain sense of community in displaying the information, the researchers say. Their analyses are based on viral DNA sequences generated and published by the National Center for Biotechnology Information, and the nonprofit Global Initiative on Sharing Avian Influenza data.

“My opinion was, since they’re publishing the sequence data openly, it makes sense for us to publish our analysis of that data openly,” said the University of Oxford’s Oliver Pybus, who with the University of Edinburgh’s Andrew Rambaut developed the site.

Analyses done so far point to the virus’ immediate origins in swine flu variants with links to avian and human strains through a process called reassortment, in which gene segments from different viruses shuffle and reassemble into new viruses.

“One important result is that the new virus derived its genes from both swine and human lineage, some of which, in turn, originated from an avian influenza pool,” said Becca Gray, a UF postdoctoral fellow working with Salemi.

Several analyses suggest that some segments of the 2009 H1N1 genome might have originated two to five years ago, and others as recently as September 2008.

“In retrospect, maybe it’s not that surprising if the strain was lost among the normal flu season in Mexico,” Pybus said. “This is suggesting that there are more people affected than those who became severely ill or died — we don’t hear about the mildly ill.”

Gray, Salemi and others are characterizing amino acid changes in the viral genome associated with the current outbreak, and investigating how those changes might affect the new H1N1 virus’ transmission rate and ability to cause disease.

While the current effort of quickly assembling and presenting data means there isn’t yet traditional peer review, the site has a self-correcting mechanism.

“If someone puts up a mistake or a silly error, they’re immediately pounced on by everyone else,” Pybus joked.

Salemi called it a sort of “worldwide peer review” because everyone can chip in with disagreements or suggestions for improvement.

“We’re not trying to pretend that what we’re putting there is what we would put in a finalized paper,” Pybus said. “We’re not explaining the results very fully, the site is very much for the cognoscenti — there’s jargon in there. Still, we’ve attempted to distill some of the most important points.”

The site employs a “wiki” — a simple program that allows contributors to upload and edit their own results. The application also lends itself to the graphics- and data-heavy computational analyses involved.

The collaborators aren’t quite sure yet how this open effort will affect their ability to publish their work in scientific journals.

“I guess it’s a bit of an experiment at the moment,” Pybus said. “It shouldn’t preclude publication later, but it’s up to the editors of the journals to make that decision. As an experiment, I hope it works out.”