Skip to Content
Biotechnology and health

“We never created a supervirus.” Ralph Baric explains gain-of-function research.

The work has helped in the development of mRNA coronavirus vaccines and the first approved covid drug. 

portrait of Ralph Baric
Ralph Baric of the University of North Carolina has engineered coronaviruses to study their threat to humans.Megan May/UNC Research

In May, the longtime coronavirus researcher Ralph Baric found himself at the center of the swirling debate over gain-of-function research, in which scientists engineer new properties into existing viruses. And during a congressional hearing, Senator Rand Paul of Kentucky implied that the National Institutes of Health had been funding such research at both the Wuhan Institute of Virology and Baric’s University of North Carolina lab, and that the two labs had been collaborating to make “superviruses.” 

Baric released a statement clarifying that according to the NIH, the research in question did not qualify as gain-of-function, none of the SARS-like coronaviruses he’d used in the experiments were closely related to SARS-CoV-2 (the original virus behind the covid pandemic), and his collaboration with the Wuhan Institute of Virology had been minimal.

Yet that did little to quell questions about the role Baric’s research may have played in furthering scientists’ ability to modify coronaviruses in potentially dangerous ways. Such questions have dogged Baric since 2014, when he became the reluctant spokesperson for gain-of-function research after the NIH declared a moratorium on such experiments until their safety could be assessed, temporarily halting his work.

Baric believes such research is essential to the development of vaccines and other countermeasures against emerging viruses, a project he has been engaged in for more than 20 years. That work has made him the country’s foremost expert on coronaviruses, and his high-security UNC lab has been a center of the US response to the pandemic, testing numerous drug candidates for other labs that lack the biosafety clearance or the expertise.

His research laid the groundwork for the first approved anti-covid drug and helped speed the development of the mRNA vaccines that have proved so pivotal. Recently, his lab announced the creation of the world’s first pan-coronavirus mRNA vaccine.

Yet Baric also pioneered the reverse-genetics techniques that have allowed other researchers, including those at the Wuhan Institute of Virology, to engineer viruses with altered functions. Some scientists fear that the technique, which allow coronaviruses to be recreated from their genetic code, could engender a future pandemic, and other critics, like Senator Paul, imply they might have led to the creation or release of SARS-CoV-2.

MIT Technology Review recently asked Baric to explain what constitutes a gain-of-function experiment, why such research exists, and whether it could have played any role in the pandemic. The interview has been edited and shortened for clarity.

Q: Now that Rand Paul has announced on the floor of the Senate that you’re creating superviruses and performing gain-of-function experiments, this seems like a good time to talk about your work.

Ralph Baric: Well, let me start off by saying that we’ve never created a supervirus. That’s a figment of his imagination and obviously being used for political advancement. Unfortunately, the way social media works today, this fabrication will be repeated many times.

How do you define gain-of-function research?

Human beings have practiced gain-of-function for the last 2,000 years, mostly in plants, where farmers would always save the largest seeds from the healthiest plants to replant the following year. The reason we can manage to have 7 billion people here on the planet is basically through direct or indirect genetic engineering through gain-of-function research. The simple definition of gain-of-function research is the introduction of a mutation than enhances a gene’s function or property—a process used commonly in genetic, biologic, and microbiologic research.

In virology, historically, attenuated vaccines were generated by gain-of-function studies, which took human virus pathogens and adapted them for improved growth in cell culture, which reduced virus virulence in the natural human host.

So gain-of-function has been used in virology and microbiology for decades as a part of the scientific method. But that classic definition and purpose changed in 2011 and 2012, when researchers in Wisconsin and the Netherlands were funded to do gain-of-function research on avian flu transmissibility.

Those were the experiments that took H5N1, which had a high mortality rate in humans but low transmissibility, and made it highly transmissible through respiratory avenues.

The NIH, the FDA, the CDC, and the WHO all held meetings to identify the critical topics in influenza research that were least understood. What information and insight would better prepare us for flu pandemics that emerge from animal reservoirs in the future? The number-one conclusion was that we needed to understand the genetics and biology of flu emergence and transmission.

In response, the NIH called for proposals. Two researchers responded and were funded, and they discovered genetic changes that regulated H5N1 transmissibility in ferrets.

After that, they were labeled as rogue scientists, and gain-of-function was defined in negative terms. But in fact, they were working within the confines of the global health community’s interests.

Then again, the other side argues that regardless of how safe your BSL-3 or BSL-4 research infrastructure is, human beings are not infallible. [Pathogen labs are assigned a biosafety level rating of 1 to 4, with 4 being the highest.] They make mistakes, even in high-containment facilities. Consequently, the risks may outweigh the benefits of the experiment. Both sides of the argument have justified concerns and points of view.

In addition to concerns over a lab escape, there were also concerns about whether the knowledge of how to do such experiments might fall into the wrong hands.

That’s certainly part of the issue. And there was a fair amount of debate about whether that information [about genetic changes associated with flu transmission] should be made public. There are two or three instances in the virology literature of papers that are a potential concern.

Some consider my 2015 paper in this light, although after consultation with the NIH and the journal, we purposely did not provide the genetic sequence of the chimera in the original publication. Thus, our exact method remained obscure.

[Baric is referring to a 2015 collaboration with Zhengli Shi of the Wuhan Institute of Virology, or WIV, in China, which created a so-called chimera by combining the “spike” gene from a new bat virus with the backbone of a second virus. The spike gene determines how well a virus attaches to human cells. A detailed discussion of the research to test novel spike genes appears here.]

However, the sequence was repeatedly requested after the covid-19 pandemic emerged, and so after discussion with the NIH and the journal, it was provided to the community. Those who analyzed these sequences stated that it was very different from SARS-CoV-2.

How did that chimeric work on coronaviruses begin?

Around 2012 or 2013, I heard Dr. Shi present at a meeting. [Shi’s team had recently discovered two new coronaviruses in a bat cave, which they named SHC014 and WIV1.] We talked after the meeting. I asked her whether she’d be willing to make the sequences to either the SHC014 or the WIV1 spike available after she published.

And she was gracious enough to send us those sequences almost immediately—in fact, before she’d published. That was her major contribution to the paper. And when a colleague gives you sequences beforehand, coauthorship on the paper is appropriate.

That was the basis of that collaboration. We never provided the chimeric virus sequence, clones, or viruses to researchers at the WIV; and Dr. Shi, or members of her research team, never worked in our laboratory at UNC. No one from my group has worked in WIV laboratories.

And you had developed a reverse-genetics technique that allowed you to synthesize those viruses from the genetic sequence alone?

Yes, but at the time, DNA synthesis costs were expensive—around a dollar per base [one letter of DNA]. So synthesizing a coronavirus genome could cost $30,000. And we only had the spike sequence. Synthesizing just the 4,000-nucleotide spike gene cost $4,000. So we introduced the authentic SHC014 spike into a replication-competent backbone: a mouse-adapted strain of SARS. The virus was viable, and we discovered that it could replicate in human cells.

So is that gain-of-function research? Well, the SARS coronavirus parental strain could replicate quite efficiently in primary human cells. The chimera could also program infection of human cells, but not better than the parental virus. So we didn’t gain any function—rather, we retained function. Moreover, the chimera was attenuated in mice as compared to the parental mouse-adapted virus, so this would be considered a loss of function.

One of the knocks against gain-of-function research—including this research—is that the work has little practical value. Would you agree?

Well, by 2016, using chimeras and reverse genetics, we had identified enough high-risk SARS-like coronaviruses to be able to test and identify drugs that have broad-based activity against coronaviruses. We identified remdesivir as the first broad-based antiviral drug that worked against all known coronaviruses, and published on it in 2017. It immediately was entered into human trials and became the first FDA-approved drug for treating covid-19 infections globally. A second drug, called EIDD-2801, or molnupiravir, was also shown to be effective against all known coronaviruses prior to the 2020 pandemic, and then shown to work against SARS-CoV-2 by March 2020.

Consequently, I disagree. I would ask critics if they had identified any broad-spectrum coronavirus drugs prior to the pandemic. Can they point to papers from their laboratories documenting a strategic approach to develop effective pan-coronavirus drugs that turned out to be effective against an unknown emerging pandemic virus?

Unfortunately, remdesivir could only be delivered by intravenous injection. We were moving toward an oral-based delivery formulation, but the covid-19 pandemic emerged. I really wish we’d had an oral-based drug early on. That’s the game-changer that would help people infected in the developing world, as well as citizens in the US.

Molnupiravir is an oral medication, and phase 3 trials demonstrate rapid control of viral infection. It’s been considered for emergency-use authorization in India.

Finally, the work also supported federal policy decisions that prioritized basic and applied research on coronaviruses.

What about vaccines?

Around 2018 to 2019, the Vaccine Research Center at NIH contacted us to begin testing a messenger-RNA-based vaccine against MERS-CoV [a coronavirus that sometimes spreads from camels to humans]. MERS-CoV has been an ongoing problem since 2012, with a 35% mortality rate, so it has real global-health-threat potential.

By early 2020, we had a tremendous amount of data showing that in the mouse model that we had developed, these mRNA spike vaccines were really efficacious in protecting against lethal MERS-CoV infection. If designed against the original 2003 SARS strain, it was also very effective. So I think it was a no-brainer for NIH to consider mRNA-based vaccines as a safe and robust platform against SARS-CoV-2 and to give them a high priority moving forward.

Most recently, we published a paper showing that multiplexed, chimeric spike mRNA vaccines protect against all known SARS-like virus infections in mice. Global efforts to develop pan-sarbecoronavirus vaccines [sarbecoronavirus is the subgenus to which SARS and SARS-CoV-2 belong] will require us to make viruses like those described in the 2015 paper.

So I would argue that anyone saying there was no justification to do the work in 2015 is simply not acknowledging the infrastructure that contributed to therapeutics and vaccines for covid-19 and future coronaviruses.

The work only has value if the benefits outweigh the risks. Are there safety standards that should be applied to minimize those risks?

Certainly. We do everything at BSL-3 plus. The minimum requirements at BSL-3 would be an N95 mask, eye protection, gloves, and a lab coat, but we actually wear impervious Tyvek suits, aprons, and booties and are double-gloved. Our personnel wear hoods with PAPRs [powered air-purifying respirators] that supply HEPA-filtered air to the worker. So not only are we doing all research in a biological safety cabinet, but we also perform the research in a negative-pressure containment facility, which has lots of redundant features and backups, and each worker is encased in their own private personal containment suit.

Another thing we do is to run emergency drills with local first responders. We also work with the local hospital. With many laboratory infections, there’s actually no known event that caused that infection to occur. And people get sick, right? You have to have medical surveillance plans in place to rapidly quarantine people at home, to make sure they have masks and communicate regularly with a doctor on campus.

Is all that standard for other facilities in the US and internationally?

No, I don’t think so. Different places have different levels of BSL-3 containment operations, standard operating procedures, and protective gear. Some of it is dependent on how deep your pockets are and the pathogens studied in the facility. An N95 is a lot cheaper than a PAPR.

Internationally, the US has no say over what biological safety conditions are used in China or any other sovereign nation to conduct research on viruses, be they coronaviruses or Nipah, Hendra, or Ebola.

The Wuhan Institute of Virology was making chimeric coronaviruses, using techniques similar to yours, right?

Let me make it clear that we never sent any of our molecular clones or any chimeric viruses to China. They developed their own molecular clone, based on WIV1, which is a bat coronavirus. And into that backbone they shuffled in the spike genes of other bat coronaviruses, to learn how well the spike genes of these strains can promote infection in human cells.

Would you call that gain-of-function?

A committee at NIH makes determinations of gain-of-function research. The gain-of-function rules are focused on viruses of pandemic potential and experiments that intend to enhance the transmissibility or pathogenesis of SARS, MERS, and avian flu strains in humans. WIV1 is approximately 10% different from SARS. Some argue that “SARS coronavirus” by definition covers anything in the sarbecoronavirus genus. By this definition, the Chinese might be doing gain-of-function experiments, depending on how the chimera behaves. Others argue that SARS and WIV1 are different, and as such the experiments would be exempt. Certainly, the CDC considers SARS and WIV1 to be different viruses. Only the SARS coronavirus from 2003 is a select agent. Ultimately, a committee at the NIH is the final arbiter and makes the decision about what is or is not a gain-of-function experiment.

Definitions aside, we know they were doing the work in BSL-2 conditions, which is a much lower safety level than your BSL-3 plus.

Historically, the Chinese have done a lot of their bat coronavirus research under BSL-2 conditions. Obviously, the safety standards of BSL-2 are different than BSL-3, and lab-acquired infections occur much more frequently at BSL-2. There is also much less oversight at BSL-2.

This year, a joint commission of the World Health Organization and China said it was extremely unlikely that a lab accident had caused SARS-CoV-2. But you later signed a letter with other scientists calling for a thorough investigation of all possible causes. Why was that?

One of the reasons I signed the letter in Science was that the WHO report didn’t really discuss how work was done in the WIV laboratory, or what data the expert panel reviewed to come to the conclusion that it was “very unlikely” that a laboratory escape or infection was the cause of the pandemic.

There must be some recognition that a laboratory infection could have occurred under BSL-2 operating conditions. Some unknown viruses pooled from guano or oral swabs might replicate or recombine with others, so you could get new strains with unique and unpredictable biological features.

And if all this research is being performed at BSL-2, then there are questions that need to be addressed. What are the standard operating procedures in the BSL-2? What are the training records of the staff? What is the history of potential exposure events in the lab, and how were they reviewed and resolved? What are the biosafety procedures designed to prevent potential exposure events?

Living in a community, workers will be infected with pathogens from the community. Respiratory infections occur frequently. No one is exempt. What are the biosafety procedures used to deal with these complications? Do they quarantine workers who develop fevers? Do they continue to work in the lab or are they quarantined at home with N95 masks? What procedures are in place to protect the community or local hospitals if an exposed person becomes ill? Do they use mass transit?

This is just a handful of the questions that should have been reviewed in the WHO document, providing actionable evidence regarding the likelihood of a laboratory-acquired-infection origin.

Should they have been doing such experiments in a BSL-2 lab?

I would not. However, I don’t set the standard for the US or any other country. There’s definitely some risk associated with these and other SARS-like bat viruses that can enter human cells.

We also know that people who live near bat hibernacula [bat caves] have tested positive for antibodies against SARS-like bat viruses, so some of these viruses clearly can infect humans. While we have no idea whether they could actually cause severe disease or transmit from person to person, you want to err on the side of increased caution when working with these pathogens.

As a sovereign nation, China decides their own biological safety conditions and procedures for research, but they should also be held accountable for those decisions, just like any other nation that conducts high-containment biological research. As other nations develop BSL-3 facilities and begin to conduct high-containment research, each will have to make fundamental decisions about what kind of containment they use for different viruses and bacteria, along with the underlying biosafety procedures.

This is serious stuff. Global standards need to exist, especially for understudied emerging viruses. If you study hundreds of different bat viruses at BSL-2, your luck may eventually run out.

Do you think their luck ran out?

The possibility of accidental escape still remains and cannot be excluded, so further investigation and transparency is critical, but I personally feel that SARS-CoV-2 is a natural pathogen that emerged from wildlife. Its closest relatives are bat strains. Historical precedent argues that all other human coronaviruses emerged from animals. No matter how many bat viruses are at the WIV, nature has many, many more.

At this time, there’s really no strong and actionable data that argues that the virus was engineered and escaped containment. As the pathogenesis of SARS-CoV-2 is so complex, the thought that anybody could engineer it is almost ludicrous.

When you think about the diversity of SARS-related strains that exist in nature, it’s not hard to imagine a strain that would have the complex and unpredictable biological features of SARS-CoV-2. As scientists, we tend to do experiments, read the literature, and then think we understand how nature works. We make definitive statements regarding how coronaviruses are supposed to emerge from animal reservoirs, based on one or two examples. But nature has many secrets, and our understanding is limited. Or as they said in Game of Thrones, “You know nothing, Jon Snow.”

In addition to the WIV and you, are other groups doing coronavirus engineering?

Before covid-19, there were probably three to four main groups globally. That’s changed dramatically. Now the number of labs doing coronavirus genetics is likely three or four times higher and continuing to increase. That proliferation is unsettling, because it allows many inexperienced groups, globally, to make decisions about building and isolating chimeras or natural zoonotic [viruses].

By “inexperienced,” I mean that they are applying previous discoveries and approaches in the coronavirus field, but perhaps with less respect for the inherent risk posed by this group of pathogens.

People are making chimeras right now for the variants of concern, and each of those variants is providing new insights into human transmissibility and pathogenesis.

So the virus itself is contributing to gain-of-function knowledge?

The virus is a master at finding better ways to outcompete its ancestors in humans. And each of these successful SARS-CoV-2 variants outcompetes the old variants and reveals the underlying genetics that regulate increased transmissibility and/or pathogenesis. And that information is being learned in a real-time setting and in humans, as compared to the avian-flu-transmission scenario, which was conducted under controlled artificial conditions in ferrets. I would argue that the real-time knowledge is more relevant and perhaps more unsettling than the research conducted in animal models under high containment.

Given our scientific capabilities today, every new emerging virus that causes an outbreak in the future can be studied at this level of granularity. That is unprecedented. Each could provide a classic recipe for potential dual-use applications in other strains. [Dual-use biological research is that which can be used to develop both therapeutics and bioweapons.]

Anything else about this that keeps you up at night?

The number of zoonotic coronaviruses that are poised to jump species is a major concern. That’s not going away.

Also, the biology of this virus is such that its virulence will most likely continue to increase rather than decrease, at least in the short term.

Why is that?

The transmission events occur early, while the most severe disease occurs late, after the virus is being cleared from the body. That means transmission and severe disease and death are partially uncoupled, biologically. Consequently, it doesn’t hurt the virus to increase its virulence.

If you are one of the people waiting to get the vaccine, your risk is going up with each new variant. These variants are dangerous. They want to reproduce and spread and show increased pathogenesis, even in younger adults. They have little concern for you or your family’s health and welfare, so get vaccinated.

That is the saddest thing about the pandemic. For an effective public health response, you need to respond as a national and global community with one voice. You must believe in the power of public health and public health procedures. Politics has no place in a pandemic, but that is what we ended up with—politically inspired mixed messaging.

How did that work out for America? Did we get diagnostics online quickly? No! Did we use the two-to-three-month lead time to stock hospitals with PPE or respirators? No. Rather, Americans received the message that the virus wasn’t dangerous, that it would go away or that the summer heat would destroy it. We heard rumors that mask wearing was detrimental, or that unproven drugs were miracle cures.

Some say that the true tragedy is the hundreds of thousands of Americans who didn’t need to die [but did] because the greatest nation in the world did not respond to a pandemic in a unified, science-based manner. Taiwan responded with a unified public health response and had only handfuls of cases and few deaths. The US led the world in deaths and numbers of cases. Why are the failures leading to the deaths of hundreds of thousands of Americans not the subject of rigorous investigation?

Deep Dive

Biotechnology and health

Scientists are finding signals of long covid in blood. They could lead to new treatments.

Faults in a certain part of the immune system might be at the root of some long covid cases, new research suggests.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

The next generation of mRNA vaccines is on its way

Adding a photocopier gene to mRNA vaccines could make them last longer and curb side effects.

Ready, set, grow: These are the biotech plants you can buy now

For $73, I bought genetically modified tomato seeds and a glowing petunia.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.