The H7N9 bird flu virus has influenza scientists on edge, due to an unexpected surge of human infections — hundreds of cases — caused by the virus this spring.

Some new scientific findings aren’t likely to ratchet down those concerns.

Scientists at the Scripps Institute in La Jolla, Calif., reported Thursday that the accumulation of several mutations in the main gene on the virus’s surface may be able to give H7N9 the ability to spread like human flu viruses do, passing from person to person through coughing and sneezing. The study, funded by the National Institutes of Health, was published in the journal PLOS Pathogens.

Currently H7N9 is a poultry virus, infecting chickens and some other birds in China. Nearly 1,500 people are known to have contracted the virus from contact with chickens since the virus first emerged in 2013, but to date H7N9 hasn’t spread easily among humans.

Read more: As bird flu spreads, US concludes its vaccine doesn’t provide adequate protection

Other scientists lauded the work, but everyone, including the scientists behind the new research, cautioned that just because something can be done in a laboratory doesn’t mean it will come to pass in nature.

“We take these kinds of things seriously. But historically we know that it is never as simple as that,” said Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases.

“So we’ve got to be careful that we separate the elegance of the science and the likelihood or not … and [what] the ultimate effect of something like this would be. Because it is a very complicated issue. And just to reduce it to one mutation or two mutations and then all of a sudden you’re in the middle of a pandemic is a bit of an overreach.”

Fauci wasn’t suggesting the Scripps team made that claim; it didn’t.

Rather, the team found that two different combinations of mutations — each made up of three changes — could in theory profoundly alter the H7N9 virus.

Flu viruses attach to receptors found on the cells of their intended victims. Bird flu viruses attach to one type of receptor. The cells found in the human upper respiratory tract — where a flu infection takes hold — are mostly lined with a different type of receptor, which explains why these viruses don’t easily infect people.

The Scripps team, however, found that by tweaking the genetic code of H7N9’s hemagglutinin gene, they could change the receptors the virus latched on to, from the bird type to the human version.

The proteins made by the two modified hemagglutinin genes attached as well to human receptors as the human virus H1N1, which caused the 2009 flu pandemic, noted Ron Fouchier, a leading influenza virologist from Erasmus Medical Center in Rotterdam, the Netherlands.

Fouchier, who has done work trying to see how H5N1 bird flu viruses could adapt to infect people, was not involved in this study.

The Scripps researchers figured out how to tweak the gene by applying what is known about flu viruses that have successfully been transmitted among people over the last half-century, as well as by studying the crystal structure of the virus. In essence, it was like custom designing a key to fit a lock.

They tried a number of combinations. Two in particular bound very well to the receptors found in human airways.

“Two combinations looked really very good. … Or really bad,” said Jim Paulson, senior author of the paper and co-chairman of Scripps’s department of molecular biology.

His team did the work, though, using only one gene of the eight-gene virus, testing the protein it generates on specially prepared glass slides.

They and others would like to see what would happen if they made these changes to a whole H7N9 virus and then tested it in ferrets, animals which serve as a proxy for people in flu research. Bird flu viruses don’t spread easily from ferret to ferret; if a modified H7N9 virus did, that would suggest it might do the same in people.

But there has been a moratorium in the United States — and among scientists elsewhere funded by U.S. government money — against doing this type of research, known as “gain of function” work or “dual use research of concern.” The latter term relates to research that might be used by malevolent actors to produce a pathogen that could be unleashed as a weapon of bioterrorism.

The moratorium resulted from alarm that arose when Fouchier and a separate team of researchers at the University of Wisconsin, Madison, tried in 2012 to publish studies that showed mutations that would be needed to the H5N1 hemagglutinin that could render transmissible among people. Those papers were eventually published but the debate over such work was and remains heated.

Fouchier said testing H7N9 viruses including the mutations identified by the Scripps team is at the top of the list of studies he plans to undertake if he can get approval for the work.

“These are the absolute critical viruses to do this with. These are a continuous threat,” he told STAT.

A new U.S. government policy for this type of work — renamed “pathogens of pandemic potential care and oversight” — is in the works. But the funding pause for such research on influenza and a couple of other types of viruses remains in effect, Ryan Bayha, of the National Institute of Health’s Office of Science Policy, said in an email.

Policy guidance from the White House Office of Science and Technology Policy issued in early January requires the department to conduct independent reviews of grant applications in which this kind of research is proposed. Bayha said the department is still setting up that review process; he could not say when the work is expected to be finished.

For now, researchers who want to test these two combinations of mutations must wait. And that means no one can say for sure if these mutations — if they were to evolve in the real world — would turn H7N9 into a virus that targets people.

“People believe that it’s important and have actually shown in some cases that it is important,” Paulson said. “But until we actually test it, we’re really not sure.”

Republished with permission from STAT. This article originally appeared on June 15, 2017