Despite Chinese obstruction and denial, the world is coming round to the idea that the origin of the pandemic might be traced to a Wuhan laboratory. Little wonder — the clues have been there all along
By Matt Ridley
It is not clear why the US government chose this week to drop not one but two bombshell announcements that a laboratory accident might have caused the covid pandemic. Both the Department of Energy (through a leak) and the FBI (through an interview) announced they are leaning towards a lab leak but neither divulged the evidence on which it bases its conclusion. The Senate has now unanimously called for the evidence to be declassified.
It looks like the Energy Department’s change of mind on this issue happened some time ago but only came to light this week, most likely for domestic political reasons. The Republican controlled House of Representatives begins hearings next week and may use the power to subpoena witnesses, which the Democrats never did. There are US-based scientists who collaborated closely with and funded the Wuhan Institute of Virology (WIV) over many years, including especially Peter Daszak of the EcoHealth Alliance, who have so far declined to testify.
One thing is certain: there is no smoking gun that implicates a lab. No evidence has emerged that this precise virus was in the WIV or any other laboratory in Wuhan before the pandemic began. But then there is no smoking gun that implicates a natural spillover either. No animal infected with this precise virus before the pandemic has been found in a market, a farm or in transit, in Wuhan or anywhere in China.
This is pretty astonishing more than three years after the pandemic began. In previous outbreaks it has usually proved possible to indict either a lab leak or a food chain within weeks or months. When a brucellosis outbreak affected more than 10,000 people in Lanzhou in late 2019, the authorities quickly traced the source to the use of expired disinfectants in a vaccine production facility.
When SARS broke out in Guangzhou in 2002, within weeks it was known that early human cases were mostly food handlers or animal vendors in markets, and that infected animals – mainly palm civets – were on sale in markets. Within two years scientists had found similar viruses in horseshoe bats, though it then took them several years to find the very closest relatives of the SARS virus in the Shitou cave in Yunnan.
The Seafood Market
Many of the early cases of covid had a direct connection to the Huanan seafood market in Wuhan, so the Chinese authorities assumed that the outbreak was a repeat of SARS. Disastrously they insisted for the first two weeks of January 2020 that it could only be caught from animals, even though by then healthcare workers were catching it from patients. Yet they tested 18 species of animal in the market and found none infected. So in May 2020, Dr George Gao, head of the Chinese Centers for Disease Control, announced that he thought the market was not the source: “At first, we assumed the seafood market might have the virus, but now the market is more like a victim.”
Dr Gao repeated that opinion last year in a detailed paper around the same time that a group of American scientists led by Dr Michael Worobey of the University of Arizona published two papers in Science which disagreed and claimed to have good evidence that the market was indeed the source. Bizarrely, therefore, by 2022 American scientists were arguing more strongly for the market hypothesis than Chinese scientists, who preferred the theory that the virus arrived in China from overseas, perhaps on frozen food.
This implausible suggestion was briefly and ingloriously endorsed by the World Health Organisation at a notorious press conference in February 2021. The WHO’s superficial investigation in early 2021 effectively played into the Chinese government’s hands by giving an illusion of investigating. Its efforts continue, in theory, but with little Chinese cooperation and no new results.
The first Worobey paper showed that of the 174 early cases reported to the WHO with onset in 2019, even those not known to have visited the market did live near it. A “heat map” of their addresses was centred on the market. Given that traces of the virus were found in the market including near where live animals had been sold – on surfaces and in sewage, but not in animal carcasses – the paper argued that it was likely an infected animal had been missed by the authorities when they tested samples in the market.
But critics soon cast doubt on these conclusions. The biggest problem is circular reasoning: until 18th January, doctors in Wuhan could only send a patient for testing for the novel coronavirus if they were linked to the market or to a market case. Also, peer reviewed Chinese papers plus a leaked document from the authorities all show, as the New Zealand based investigator Gilles Demaneuf reported, that there were about 260 cases in 2019, not 174. Moreover, there is a hospital and a lab (the Wuhan Centers for Disease Control, which had been harvesting bat viruses) very near the market, so also near the centre of the heat map.
Worobey’s other paper with Jonathan Pekar argued that since one case of a second human strain of the virus was found in the market, there must have been two different spillovers from separate animals. This makes the failure to find a single infected animal or human case on a farm or in the supply chain of the market all the more surprising. Divergence of the strains during prior low-level circulation among people is an alternative explanation.
The Worobey papers did not as the authors hoped kill off the lab leak theory. Those (like me) who insisted that neither the market nor the lab could yet be ruled out have stuck to that view.
The Horseshoe Bats
We know for sure that SARS-CoV-2 is a natural bat virus. Even if it has been subject to genetic manipulation, most of its genome is similar to that of other SARS-like beta coronaviruses. The natural habitat of all such “sarbecoviruses” is inside the intestines of one genus of bat, the horseshoe bats, a fact that was discovered in the wake of the first SARS epidemic in 2002-3 by Dr Shi Zhengli of the Wuhan Institute of Virology.
Viruses with very similar genomes to SARS-CoV-2 have turned up in horseshoe bats in Yunnan and Laos, while more distant relatives have been found in Cambodia, Thailand, eastern China and Japan. But none has shown up near Wuhan, despite over ten thousand horseshoe bats having been sampled in the area by the Wuhan CDC. When Dr Shi first heard about the outbreak, she told a journalist: “I had never expected this kind of thing to happen in Wuhan”.
The central question is how a virus travelled from a bat in Yunnan or Laos to Wuhan, more than a thousand miles to the north-east. Horseshoe bats don’t migrate such long distances, so there are four plausible carriers: an infected animal on its way to market, an infected person who visited a bat cave, an infected scientist who visited a bat cave looking for viruses, or a sample taken from a bat to a laboratory.
A clue emerged in January 2020. Dr Shi published the genome of the new virus in Nature magazine and in her article she revealed that a 96.2% similar virus had been “previously detected” in a horseshoe bat from Yunnan, but did not say where or when and gave the virus a previously unused name, RaTG13. It took those of us interested in the topic two months to find out that RaTG13 came from a mineshaft where three miners had died of a mysterious pneumonia in 2012. This incident had caused widespread alarm that SARS might have jumped directly from bats into people.
RaTG13 had been found in 2013 in one of seven expeditions mounted by the WIV to the mineshaft over three years. We later found out, no thanks to Dr Shi, that eight other similar viruses had been brought back to Wuhan from the same site, and that RaTG13 had been genetically sequenced in Dr Shi’s lab in 2018, contradicting assurances from Dr Daszak that it had been resting quietly in the lab’s freezer in the intervening years.
So not only had a bat coronavirus pandemic started in the city with the biggest bat coronavirus research lab in the world but the nine closest cousins of the virus causing the pandemic had at the time been in that very lab and one of them had recently been subject to laboratory work. Yet none of these cousins was close enough to be the actual progenitor of the human outbreak and Dr Shi insisted she had never had SARS-CoV-2 itself in her lab. An argument against the lab leak is that western scientists trawling through deep sequencing data can sometimes find traces in databases of what other labs are working on; so far nothing on this virus before the outbreak has surfaced.
The Laos Viruses
RaTG13’s importance took a knock in 2021 when French scientists working in Laos reported that they had found a bat virus even more closely related to SARS-CoV-2. Called Banal-52, this virus is 96.8% the same as SARS-CoV-2. There are two reasons, however, that this discovery does not exonerate the lab. First, the new virus is actually slightly less similar over most of its genome than RaTG13, except in the spike gene where it is much more similar. The spike gene is the very bit that scientists have sometimes altered. Second, we know from genetic databases that the EcoHealth Alliance has sampled viruses in Laos and sent them to Wuhan.
Wouldn’t we know of all the viruses sent to Wuhan? No. The WIV refuses to this day to share the contents of its database of bat viruses, even though the purpose of assembling the database was to alert the world to risky viruses that might cause pandemics. Sharing that database would rapidly exonerate the WIV if it contains nothing close to SARS-CoV-2. Dr Daszak has published many of the viruses found before 2016 but few since then. When asked whether he had asked Dr Shi for the database when in Wuhan as part of the WHO investigation, he replied: “We did not ask to see the data…We do basically know what’s in those databases.”
Dr Shi has insisted that she has published all the viruses she was working on, but this is clearly not true: the eight other viruses from the mineshaft went unmentioned until they showed up apparently by accident in a genetic database and in a thesis downloaded from a private Wuhan website by an ingenious Indian sleuth, Jeet Ray.
The WIV did not just hoard bat sarbecoviruses; it experimented on them. By swapping spike genes between bat viruses they sometimes increased the infectivity of the viruses 10,000-fold in mice with human genes. Some of these experiments were done at inappropriately low biosafety levels. But again, none of the published experiments used a virus that could have directly caused covid. Were there unpublished ones?
The Furin Cleavage Site
The reason we had a pandemic and not a brief epidemic is that SARS-CoV-2 contains a unique genetic feature found in no other sarbecovirus, which makes it much more infectious. Called a furin cleavage site, it is a short insertion of genetic material at a critical position in the spike gene, which has the effect of opening up the spike protein in a special way. In early 2020 there were two starkly different reactions to this discovery.
The strangest was Dr Shi’s. In her paper describing the genome of the new virus she drew attention to several features of the spike gene but said nothing about the furin cleavage site. As my co-author Dr Alina Chan puts it, this was like describing a unicorn but not mentioning the horn.
The other reaction was from a handful of western virologists who worried privately that the furin cleavage site suggested a lab leak as the source. This was because they knew that inserting furin cleavage sites into viruses has become a bit of a hobby in certain virology labs in recent years, including in Wuhan and including Dr Shi, who co authored a paper in 2015 reporting the engineering of a cleavage site in the spike of a MERS-like virus.
In an online meeting on Saturday 1st February 2020 a dozen senior scientists discussed their worries. A week later a Chinese university announced they had found a pangolin with a virus 99% similar to SARS-CoV-2 and the scientists speculated in their emails that this new pangolin virus might prove to have a natural furin cleavage site, which would lay their concerns to rest. It does not, and it is only 90% the same, not 99%, so pangolins are a red herring. Still, in the paper they drafted these scientists asserted that a sarbecovirus would soon turn up with a natural furin cleavage site. Three years later none has.
What has turned up is a document authored by Dr Daszak, Dr Shi and others in 2018 requesting $14 million from the US government to do a series of experiments including one in which they would insert a furin cleavage site into a sarbecovirus for the first time. That grant application, called the Defuse proposal, is about as close as you can get to a recipe for creating SARS-CoV-2. The proposal was turned down but might have been funded by the Chinese government instead.
Where does the debate now stand? Opinion polls show that even before this week almost twice as many Americans thought the pandemic began with a lab leak as thought it began naturally. Many media outlets continue to state as fact that the majority of scientists thinks it began in the market but not only are there no reliable polls of scientists; many scientists say it could be career suicide even to show interest in the lab leak.
To believe in the market theory you have to assume that the Chinese authorities missed or covered up some infected animals. To believe in the lab theory you have to assume that they missed or covered up an experiment with a precursor of SARS-CoV-2. One of those is true.
Matt Ridley is the co-author, with Alina Chan, of Viral: the search for the origin of covid-19.