Gain-of-function research on viruses justifies itself with a scientific error

We still don’t know where Covid-19 started, although we’re pretty sure it started in or near the city of Wuhan, China. The leading theories are that it started either in the Huanan Wholesale Seafood Market (in Wuhan), a live animal food market, or in the Wuhan Institute of Virology (WIV), a large virus research center in another part of the city.

We might never know, because we’d need access to all of the viruses being studied at WIV in late 2019, and those viruses might not even exist any longer.

I’ve been on the fence about this question since the pandemic started (as I wrote here and here and here), in part because we just don’t have enough data. However, I’m now starting to lean more strongly towards the hypothesis that the Covid-19 virus started in the Wuhan Institute of Virology. I just listened to the interview that Sam Harris did with science journalist Matt Ridley and biologist Alina Chan, who together wrote an entire book on the origins of Covid-19, and the evidence they compiled is compelling.

Let’s look at a few key points.

First, the virus itself, SARS-CoV-2, almost certainly originated in bats, and those bats almost certainly came from caves in southern China, over 1000 kilometers away from Wuhan. The bats didn’t get to Wuhan on their own.

So either someone transported bats to the Huanan Seafood Market, or they transported viruses from the bats to WIV. These are our choices.

Second, WIV was doing research on coronaviruses for years. Their scientists traveled regularly to caves in southern China to find novel viruses, and they’ve acknowledged that WIV’s labs had bat viruses, including viruses related to SARS-CoV-2, before the pandemic began.

Third–and this point is under some dispute–many scientists have argued that the virus was a naturally-occurring one. However, this doesn’t make it more likely that the virus originated in the seafood market. It’s just as likely that a scientist or technician working at WIV was accidentally infected, and then went home (maybe stopping by the seafood market on the way) and started a worldwide pandemic.

Fourth, it’s hard to believe that it’s merely a coincidence that one of the top virology labs in China just happened to be located in the city where the pandemic began. WIV was not only the foremost lab in China doing work on SARS-like viruses, but they also claimed previously that they intended to do gain-of-function work to make these viruses more pathogenic.

This startling fact emerged when a 2018 grant proposal by EcoHealth Alliance, a US-based nonprofit that was working with WIV, was leaked to the press in 2021. Even though that proposal was never funded, the text describes how EcoHealth would genetically engineer new viruses, taking the spike protein from one bat coronavirus and inserting it into a different one, and then infecting mice to see what happens.

But wait, some will say: we now have peer-reviewed studies pointing to the seafood market as the epicenter of the pandemic. (I wrote about those studies back in March 2022.) However, as Alina Chan and Matt Ridley explained to Sam Harris (and in their book), the Chinese authorities in early 2020 focused all their attention on the seafood market, to the exclusion of anywhere else. They collected loads of samples from people who had been in or near the market, and very little from anywhere else. Thus we seem to have a classic case of confirmation bias: when you only look at the place where you’re convinced the virus originated, and you find some evidence, then you stop looking. We simply don’t know if the virus was anywhere else.

Now, to the main topic for today: the scientific error used to justify gain-of-function research on dangerous viruses, the error that might have led to the Covid-19 pandemic. Let me explain.

Why, one might ask, were scientists from the Wuhan Institute of Virology going out into the wild, to places where humans would not otherwise go, and bringing back deadly viruses?

This doesn’t just happen in China. The US is funding a large effort to do exactly the same thing: through a program called DEEP VZN (”deep vision,” get it?), USAID is funding scientists in the US and in Africa, Asia, and Latin America to venture into unpopulated areas of the jungle, and to find animals carrying viruses that might infect humans. They’re hoping (!) to discover 8,000 to 12,000 new viruses, and they’re particularly interested in viruses that could cause the next pandemic.

Why does anyone do this? Virus hunters believe that through these efforts, they can predict which of these viruses are destined to become the next pandemic. Furthermore, the argument goes, through gain-of-function research, virologists will be able to determine just what the new pandemic strains will look like. Then, armed with this knowledge, they can convince governments and private companies to design, manufacture, and stockpile vaccines against these viruses. This way (the argument goes) when the pandemic emerges, we’ll be ready.

At the center of this scientific strategy is an obvious error about evolution.

I’ll have to get a bit technical to explain here, so bear with me: the genome of the SARS-CoV-2 virus contains about 30,000 bases of RNA. The key protein that lets it infect human cells is called the Spike protein, which is about 1300 amino acids long and is encoded by about 3900 RNA bases. RNA has an alphabet of 4 letters (A, C, G, and U), which means that each of those positions can mutate into one of the other 3 letters. So we have almost 12,000 possible mutations that affect just one base in the Spike protein.

But 2 or more mutations can happen at once, quite easily, and this too could make the virus more infectious. How many combinations of 2 positions and 3 mutations are possible? Well, about 650,000,000.

And these aren’t the only mutations that might create a pandemic virus. So we’re supposed to believe that:

  1. gain-of-function experiments in the lab will create precisely the same mutations that would happen in the wild, and
  2. virologists can then predict, based on their experiments, that a virus is likely to cause a pandemic, and
  3. this evidence is so convincing that governments will manufacture and stockpile vaccines based on these experiments, and
  4. that this in turn will allow us to prevent the next pandemic.

Yeah, right. The evolutionary mistake is in the first point above, by the way.

Something happened in Wuhan. You might think that virologists, upon hearing about the gain-of-function research at WIV, would pause and think, oh no, we hope our colleagues’ research didn’t cause the pandemic! But instead, they closed ranks and doubled down.

In case you think I’m exaggerating, consider this: just a month ago, 156 virologists co-authored an article in the Journal of Virology that declared:

“gain-of-function research-of-concern can very clearly advance pandemic preparedness and the development of vaccines and antivirals. These tangible benefits often far outweigh the theoretical risks posed by modified viruses.”

In case that wasn’t clear enough, they assert twice more in the article that gain-of-function research will help us prepare for pandemics.

Virologists have been making this argument for years, and yet their experiments had no benefit at all–none, zero, zip–when we were finally faced with a true pandemic. Why should we believe this claim now?

Instead, it’s possible that gain-of-function research, along with the search for novel viruses in the wild, might have accidentally caused the pandemic.

Let me conclude by emphasizing that the vast majority of research on viruses and infectious diseases is incredibly important. Vaccines, antibiotics, antivirals, and other treatments have saved millions of lives, and the scientists doing this work should be applauded. Shutting down dangerous gain-of-function research–by which I mean research designed to take a virus or bacterium and make it more deadly in humans or in other animals–would only affect a tiny percentage of virologists worldwide. Let’s tell them to stop. If they can’t find something better to do, other scientists can.

Panel recommends new controls on deadly gain-of-function research. Will the government listen?

Illustration by Erik English

This past week, a government-appointed panel of scientists released a new report recommending 13 actions the U.S. government should take to control “gain-of-function” research that has the potential to create deadly new pathogens.

This has been a long time coming, but the first thing I want to point out is that this is just an advisory panel. The government hasn’t done anything yet. Let’s unpack what happened, shall we?

First, the panel is called the NSABB: the National Science Advisory Board for Biosecurity. The new report, which was at least 3 years in the making, was created in response to a decade’s worth of concerns, raised by many scientists (including me - see my previous articles here and here and here, among others), about the dangers of a specific kind of research known as gain-of-function.

What is gain-of-function (GoF) research? Well, it can include many scientific experiments, including some that are perfectly reasonable. But the term has been used most often to refer to experiments that are designed to take a virus such as influenza or SARS-CoV-2 and alter it intentionally to make it more deadly.

This seems crazy, right? Yet it’s been going on in the influenza virus research world for at least a decade, which is why many scientists have raised alarms.

The Covid-19 pandemic gave this issue much greater urgency, after suspicions arose that the Covid-19 virus, SARS-CoV-2, might have emerged (accidentally) from gain-of-function experiments at a major virology lab in Wuhan, China. (It probably didn’t, but we still don’t know for sure, as I’ve explained in previous columns.)

So back to the topic at hand: the new NSABB report. What do they recommend, and will it matter? I don’t want to go through all 13 recommendations, but overall it’s a very good start, if (and only if) the U.S. government takes them seriously and implements them all.

And the virology community is already pushing back - but first let me go into just three of the recommendations.

First, the panel recommends that the government require that all GoF research undergo federal-level review if the work is

“reasonably anticipated to enhance the transmissibility and/or virulence of any pathogen.”

Believe it or not, GoF research that does this kind of thing is going on right now, and there’s no rule saying it must be reviewed first.

Second, the panel recommends that the government only allow such research if there’s simply no better, lower-risk way to gain the same scientific insights. As they put it, scientists who want to do GoF work would have to prove that

“there are no feasible alternative methods ... that poses less risk ... and the risks are justified by the potential benefits.”

That’s a high bar to clear, but it seems eminently reasonable to insist upon it before allowing dangerous GoF research to proceed.

The panel also recommends that the new restrictions on gain-of-function research apply to all research in the U.S., regardless of whether it’s funded by the government. This is an important addition, as illustrated recently when Boston University, after being called out for dangerous gain-of-function experiments on the Covid-19 virus, claimed that they didn’t use NIH funds for this, so (they argued) they didn’t break any rules. Technically, they were correct. This recommendation will close that giant loophole.

There’s much more in the NSABB report, and my primary reaction is that (1) it’s a good start and (2) it’s not nearly enough. I’d like to see the government make a blanket statement that research that will make deadly viruses even more deadly is simply forbidden, at least for now. If someone wants an exception, they could make the case, but I’ve yet to see a good argument for these experiments.

What about that pushback from the virologists that I mentioned above? Well, in a lengthy commentary just published in the Journal of Virology, 156 virologists argue that gain-of-function research is wonderful! And it’s brought so many benefits! Just let us handle this, and don’t worry, they seem to be saying.

To make the benefits explicit, the 156 virologists include a table listing dozens of “useful examples” of gain-of-function research. Let’s look at just two of them.

Example 1: the virologists assert that experiments on a virus called M13 led to faster computers, citing a 2018 article. First, this is nonsense: no one has been a faster computer using a modified M13 virus. Second, the M13 virus is harmless to humans (it only infects bacteria), so it wouldn’t be subject to any regulations on GoF research in human pathogens.

Example 2: this one is even more outrageous. The table lists as a “benefit” an experiment that “established that H5N1 has capacity for mammalian transmissibility.” They then cite a notorious experiment from 2012 in which scientists intentionally modified a deadly bird flu virus (H5N1) in order to make it possible for the virus to be transmitted directly between mammals. This was one of the key experiments that led to the widespread alarm about GoF research in the first places. (I wrote about it back in 2013.)

So no, creating a more-deadly virus and then saying “see? look how dangerous this virus is?” is not what I’d call useful.

Clearly, the virologists who wrote this commentary do not want to see any restrictions at all on the kind of research they do. They just don’t see the need for it. Obviously, I disagree, as do many others, including many virologists who support the NSABB recommendations.

As I wrote at the beginning of this piece, the NSABB report is just a set of recommendations, and the government might not do anything. I hope that the government will implement all of them, and then go even further, and put a stop to the dangerous, sometimes reckless experiments that a very small minority of scientists are engaging in.

We need to study viruses, and we need to control infectious diseases, but we can do this without making pathogens more deadly.