The pandemic originated from a natural transmission of a virus from an animal to a human.

Zoonosis

The pandemic emerged as a result of a virus that leaked from a lab in Wuhan. 

Lab Leak

Previous coronavirus pandemics (SARS1 from China 2002, and MERS from Saudi Arabia 2012) were not infectious enough to cause a pandemic and HKU-1 which was not lethal enough.
Thus, the estimate is that the world experiences a major zoonotic bat coronavirus pandemic once every twenty years, or 5% per year.

Annual likelihood of a zoonotic pandemic

East Asia is home to a diverse distribution of bat species, which were the source for SARS in 2002, likely reaching humans through civets sold at a meat market in the Guangdong Province in southern China. The origin of another dangerous coronavirus (MERS) was in the Middle East. In order to account for the greater likelihood of a zoonotic outbreak in East Asia (the original SARS outbreak location, and where there are more wildlife-human interactions), we compare the population of Wuhan only to urban areas in other relevant East Asian countries, as Wuhan is no more likely to have a bat coronavirus zoonosis than other major East Asian cities.
<ul>
<li>In 2020, Wuhan's population stood at 12.3 million.</li>
<li>The total urban population in East Asia is 1.66 billion.</li>
<ul>
<li>This includes the urban populations of Thailand, Malaysia, Nepal, North Korea, Taiwan, Laos and Singapore, and half the urban populations of the more remote India and Sri Lanka.</li>
</ul>
<li>Multiplied by 2 because we need the virus to not die out, and the larger the city the more likely it is due to higher density.</li>
</ul>
The probability of Wuhan as an outbreak location is therefore 1.48%.

Likelihood of Wuhan as the outbreak location

To assess the probability per year of a bat coronavirus pandemic starting in Wuhan due to a zoonosis, we estimate the annual probability of such a pandemic anywhere, multiplied by the conditional probability of Wuhan being the outbreak location.

 Zoonosis

Although the DEFUSE proposal was rejected, it still reveals the research interests of WIV and the other participants.
In 2019, the study of furin cleavage sites gained prominence in coronavirology. Ralph Baric's group from the University of North Carolina investigated the spike protein cleavage of MERS viruses. Interestingly, Baric's 2020 research did not involve collaboration with Zhengli Shi and WIV, despite past joint studies on FCS in coronaviruses. This shift followed the undisclosed 2018 DEFUSE proposal by EcoHealth, WIV, and Baric, which focused on FCS in SARS-like viruses. After DEFUSE's funding fell through, <a href="https://twitter.com/ydeigin/status/1680095939757846531" target="_self">WIV and Baric may have pursued their research independently, driven by competitive dynamics</a>.
Additionally, WIV was planned to do a lot more than officially written in DEFUSE. <a href="https://usrtk.org/covid-19-origins/american-scientists-misled-pentagon-on-wuhan-research/" target="_self">Recent discovery shows it was downplayed</a> to help get DARPA approval.
The probability of WIV initiating a project similar to DEFUSE, despite its rejection, is therefore estimated at 25%.

WIV started a DEFUSE-like project

The BANAL-52 virus is identical to SARS2 in that aspect, indicating that such viruses exist in nature. Therefore, it is probable that a suitable match could be found with enough searching, giving it a 50% likelihood.

WIV screened for and optimised human ACE2 binding

Given that this work was done in BSL-2 a lab worker infection is quite likely: this is estimated to be 15% per year.

A lab worker got infected

Given infectivity of the virus, it is likely that an infected lab worker would start a pandemic in Wuhan (40%).

The infected lab worker started a pandemic in Wuhan

For a bat coronavirus pandemic to start in Wuhan due to a lab leak, the following chain of events needs to occur:
<ul>
<li>WIV started a DEFUSE-like project.</li>
<ul>
<li><a href="https://theintercept.com/2021/09/23/coronavirus-research-grant-darpa/" target="_blank">In 2018, DEFUSE, a grant proposal</a> to DARPA aimed at identifying and culturing natural sarbecoviruses that infect human cells described the insertion of human-specific cleavage sites into SARS-related bat coronaviruses. This proposed experiment, which could have led to the creation of SARS-CoV-2, was rejected. However, it is possible that WIV conducted a similar project nonetheless.</li>
</ul>
<li>WIV screened for and optimised human ACE2 binding then optimized it for human infection (e.g. added an FCS)</li>
<li>A WIV lab worker got infected by the virus</li>
<li>The infected lab worker started a pandemic in Wuhan</li>
</ul>

 Lab Leak

The starting point in Bayesian analysis is the "priors" associated with an event, representing the initial probabilities or likelihoods of different hypotheses before any specific evidence is considered. These priors are crucial as they form the starting point from which analysis begins and beliefs are updated upon reviewing new, relevant data.

What is a "starting point"?

To evaluate the starting point for an analysis, we usually begin by selecting a subset of evidence that is easy to analyze and gather statistics for. In this case, estimating priors is challenging due to the rarity of major pandemics and of lab leaks.
Additionally, rapid advancements in genetic engineering complicate the assessment of the current risk of a major pandemic caused by Gain of Function research, relative to the lack of any known cases in the past.
Therefore, we calculate an estimated annual emergence rate for each hypothesis, specifically, the probability per year of a bat coronavirus pandemic starting in Wuhan.

Starting Point

Source : <a href="https://www.nature.com/articles/s41598-021-91470-2" target="_blank">https://www.nature.com/articles/s41598-021-91470-2</a>&nbsp;
Wuhan, China has 4 wildlife markets associated with selling animals that could have acted as intermediate hosts for the virus; Huanan Seafood Market, Qiyimen Live Animal Market, Baishazhou Market, and the Dijiao Outdoor Pet Market.

There are 7 live animal shops in HSM and a total of 17 in the four Wuhan markets.

Source : <a href="https://finance.sina.com.cn/roll/20050725/0855226709.shtml" target="_blank">https://finance.sina.com.cn/roll/20050725/0855226709.shtml</a>&nbsp;

HSM is the largest seafood market in central China

Source: <a href="https://www.who.int/docs/default-source/coronaviruse/who-convened-global-study-of-origins-of-sars-cov-2-china-part-annexes.pdf" target="_blank">Joint WHO-China Global study of origins report</a>, p. 175

28% of early cases (December 2019) were exposed to HSM

For the animal hypothesis, we need to estimate the probability that HSM would be where the virus spillover occurs, as well as account for the evidence of wildlife involvement in HSM.
<ul>
<li>There are three other known markets in Wuhan that sell wildlife: Qiyimen Live Animal Market, Baishazhou Market, and the Dijiao Outdoor Pet Market. It is also possible that illegal trade occurs.</li>
<li>In the SARS1 outbreak, spillovers were repeatedly occuring in restaurants rather than markets.</li>
</ul>
Therefore, we estimate p(HSM|Wuhan,Zoonosis) = 10%, meaning that given an outbreak in Wuhan due to zoonosis, HSM has a 1 in 10 probability of forming the earliest detected cluster.

Animal hypothesis - An animal brought the virus to HSM

Two lines of reasoning point to HSM being a relatively likely location for an early cluster, given infections in Wuhan:
<ul>
<li>A number of factors unique to HSM that make it more conducive for SARS2 growth.</li>
<li>Empirical data pointing to seafood markets and facilities forming early clusters in other cases.</li>
</ul>

Human hypothesis - A human brought the virus to HSM

There are two hypotheses for why HSM formed an early cluster of Covid cases: either
<ul>
<li>An infected wild animal was brought to HSM, presumably to be sold, where it infected humans ("the animal hypothesis"); Or</li>
<li>A wild animal infected humans outside HSM, it subsequently spread throughout Wuhan, while specific conditions in HSM made it a likely location to form an early cluster it ("the human hypothesis").</li>
</ul>

Either an animal or a human brought the virus to HSM

This hypothesis is identical to the “human hypothesis” best explanation under zoonosis, we therefore use the same figures.

A WIV worker was infected, transmitting the virus to another individual within HSM. The virus then arrived at HSM, where the conducive environment accelerated the virus's growth, forming an early cluster.

HSM is a conducive environment to forming an early SARS-CoV-2 cluster

Many early cases were associated with the Huanan Seafood Market (HSM), the largest seafood market in central China and one of a handful of markets selling live animals in Wuhan.

Location (HSM)

Although it seems sarbecoviruses are unlikely to acquire an FCS, it is difficult to assess exactly by how much. We ignore this evidence to avoid double counting with the FCS features, which we will analyze later.

FCS are <a href="https://www.sciencedirect.com/science/article/pii/S1097276520302641" target="_blank">not common in coronaviruses</a>, and <a href="https://www.sciencedirect.com/science/article/pii/S2589004220308348" target="_blank">never appeared in SARS-related coronaviruses</a> before or since SARS-CoV-2. But given that we have a bat coronavirus pandemic, it is not unreasonable to expect that the new virus will have some novel feature that increases its ability to infect humans, and an FCS is relevant.

A new virus starting a pandemic is expected to have a novel feature 

The presence of an FCS in itself is expected and therefore this evidence group is ignored.

WIV <a href="https://theintercept.com/2021/09/23/coronavirus-research-grant-darpa/" target="_blank">has expressed</a> a specific interest in adding FCSs, so this is relatively expected.

WIV is expected to add FCSs

SARS-CoV-2 has a Furin Cleavage Site (FCS), an amino acid sequence that causes the protease furin to cut the virus in a way that facilitates its entry into cells.
This evidence is noteworthy because the Wuhan Institute of Virology (WIV) <a href="https://theintercept.com/2021/09/23/coronavirus-research-grant-darpa/" target="_blank">showed interest in incorporating FCSs</a> into viruses.
This evidence group only addresses the presence of an FCS, not its characteristics, which will be analyzed in the following evidence groups.

Presence of an FCS

Since no one was able to point to such an explanation by now, then a conservative estimate is that:
<ul>
<li>There is a 5% likelihood that an explanation exists in which</li>
<li>There is a 10% probability that the way the FCS would appear is a long clean insertion, given that the virus has a FCS is 0.1 (that is : p(genetic features|FCS,ZO) = 0.1)</li>
<li>Which combined is 0.5%</li>
</ul>

There is some unknown reason why an FCS specifically should emerge with a long insertion. Despite years of discussion, no suggestion has been put forward, and there are no examples of similar phenomena in other viruses. The best explanation is therefore that there is some evolutionary mechanism that we do not yet understand, which makes a long insertion less rare and more preferable.

Unknown evolutionary advantage

In the past, FCSs have been added by substitution rather than insertion, but it is not difficult to do it by insertion. We cannot be sure of the exact considerations of the lab researchers who added the FCS, such as investigating the role of proline. Therefore, we assign it a 10% probability.

Under lab leak, these genetic features are reasonable choices for an engineer as insertions are a simple, common engineering tool.
It is possible that the scientists at the WIV were interested in investigating something related to the leading proline (discussed below) and, since it was not present, they needed to insert it.

Common engineering tool, possibly to explore the role of proline

The FCS emerged in SARS2’s genome as a clean 12 nucleotide insertion, implying three unexpected features :
<ul>
<li>It is an insertion, <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8077628/" target="_blank">which are very rare.</a> The FCS could have developed through standard single nucleotide variations (SNV), which are 10,000 times more common. In fact, several coronaviruses are a few mutations away from an FCS, and there is no known advantage to obtaining the FCS through an insertion rather than an SNV.</li>
<li>It is a long insertion instead of a more common short insertion (e.g. of 3 nucleotides only)</li>
<li>It is a clean insertion, appearing in a highly conserved area of the genome, with many synonymous SNVs.</li>
</ul>
This is in contrast to the far more likely alternative of acquiring an FCS through SNVs.
<ul>
<li>For example the RmYN02 virus is only one nucleotide mutation away from turning its PAAR sequence into RAAR, which is a suboptimal cleavage site that can later develop into a fuller RARR or RRAR cleavage site.</li>
</ul>

12 nucleotides clean insertion

A study on SARS2 insertions did identify several insertions that could not be found in the SARS2 genome, but they were not fully investigated to determine the source. In any case, the sequence CGGCGG is very unique, making the insertion especially suspicious under zoonosis. It is therefore given a conditional probability of 15%.

The best explanation is that the insertion originated from other genetic material present in the cell where it occurred, presumably from the host.

The insertion is from genetic material in the cell

Since these features are a reasonable choice for an engineer, the estimated likelihood of this best explanation is 70-80%.

A genetic engineer interested in working on an FCS does not have to choose a sequence that already exists in SARS2 or any other virus. Instead, they could introduce any sequence they wanted.

The insertion was chosen by the engineer

The 12 nucleotides sequence does not appear anywhere else in the SARS2 genome, meaning it was not copied internally as insertions often happen, but instead the sequence came from some foreign source.
The two arginines are coded by the rarest codons for arginines (CGG) and the sequence CGGCGG doesn't appear in the SARS2 genome

Insertion is from a foreign source

Having random features is expected under zoonosis. However, we could have had many more redundant or suboptimal features, such as having a much longer insert with more amino acids, or having it out of frame without the possible explanation of one SNV right before the insert. As such, reasonable likelihoods for it are 40%-60%, with 40% being our impartial estimate.

The best explanation is that since evolution involves random mutations, it is therefore likely to add redundant or suboptimal features. Regarding the "out of frame" feature, the best explanation is that it is indeed out of frame since inserts in nature can happen in any location.

Evolution is expected to include random features

These explanations, while not perfect, provide reasonable explanations for the suboptimal FCS features. Determining what an engineer was looking for is difficult, so we estimate a range of 3-10% likelihood for this best explanation.

It is difficult to claim very low probability for specific genetic features under the lab leak hypothesis, as we are not aware of the specific motivations of WIV scientists.
That being said, one plausible explanation for the inclusion of the proline amino acid and the non-canonical RRAR is that WIV could have been interested in simulating a natural evolution from PAA to an FCS similar to the one observed in the Middle East Respiratory Syndrome (MERS) coronavirus:
<ul>
<li>One of the closest known bat viruses to SARS-Cov-2, RmYN02, has the sequence PAAR at the S1/S2 junction</li>
<li>If researchers at WIV encountered a virus akin to RmYN02 containing a PAA fragment, they might have hypothesized a natural evolution path from PAA to a MERS-like FCS, possibly involving PAA -&gt; PRA -&gt; PRRA transitions. In such a scenario, the insertion and the leading proline are a plausible approach, as no PAA exists at that position.</li>
</ul>
Regarding the "out of frame" feature , the best explanation is that it is actually in frame with one SNV right before the insert.

Unidentified factors

Three features are suboptimal for an FCS to work :
<ol>
<li>Leading proline : the amino acid sequence of the insert is PRRA, while theoretically RRA would suffice for engineering purposes.</li>
<li>Non-canonical FCS: the insertion creates the FCS RRAR instead of the more canonical cleavage sites RARR or RKRR.</li>
<li>“Out of frame” : the 12 nucleotides insertion could be interpreted to be out of frame (one nucleotide earlier than being in frame) or to be in frame with one SNV right before the insert.</li>
</ol>

Suboptimal features for an FCS to work

"Based on epidemiological analysis and genetic sequencing of viruses from new cases throughout 2020, there is no evidence of repeated introduction of early SARS-CoV-2 strains of potential animal origins into humans in China. There was no genetic or serological evidence for SARS-CoV-2 in a wide range of domestic and wild animals tested to date. The screening of the major livestock species was done across the country and provided no evidence for circulation of a related virus. The scale of testing in these species was such that widespread circulation is extremely unlikely. Screening of farmed wildlife was limited but did not provide conclusive evidence for the existence of circulation."
Source: <a href="https://www.who.int/docs/default-source/coronaviruse/final-joint-report_origins-studies-6-april-201.pdf" target="_blank">WHO report</a>, p. 116

The final origin report from the WHO indicated that no intermediate host was found despite extensive testing

Source: <a href="https://m-gatta.medium.com/list-animals-coronavirus-31414ca9576f" target="_blank">https://m-gatta.medium.com/list-animals-coronavirus-31414ca9576f</a>&nbsp;

Many animals were found to carry SARS2, indicating constant monitoring that should be able to identify infections related to ancestral SARS2 strains

Under zoonosis, an intermediate host for the virus likely exists, considering that the virus is not highly effective in bats. It is also expected that a major pandemic will be seriously tested, especially as finding the host is crucial to prevent more spillovers in the early days of a pandemic.
<ul>
<li>China was trying to reach zero infections at the time, which could only have been done by finding the intermediate hosts and ensuring humans don’t come into contact with them.</li>
<li>Intermediate hosts were identified in other pandemics</li>
<li>Furthermore, exponential advances in genome sequencing increase the efficiency of such searches compared to earlier cases like SARS1 .</li>
</ul>
The more spillovers there are, the more animals should be infected, and the less likely we are to not find an intermediate host, or have another spillover anywhere in the world. This is particularly relevant given the multiple outbreaks in China that were caused by imported food or contacts during the Zero COVID period - i.e. there were many new introductions, but none were from a new zoonotic spillover.
However, lockdowns and the closing of the market may have made tracking down intermediate hosts harder.
While this best explanation seems unlikely, there isn’t enough experience with similar situations to confidently give it a low prior probability, so it is estimated at 25%.

The ancestor virus spread among an animal population that lives far from humans, creating few opportunities for spillovers or for catching an infected host. Just a few of them were captured, causing a spillover in Wuhan.
The population still needs to be large enough to allow for the development of the FCS.

Small group of isolated infected animals

Under lab leak, there is no intermediate host, so it should not be found, and there shouldn’t be any additional spillovers. Therefore the conditional probability of this evidence group is 100%.

No intermediate host or additional spillover is expected

No animal was found to be infected with a virus related to an ancestral strain of SARS2, i.e. which is genetically closer to bat coronavirus than the early SARS2 strains.
No SARS2 outbreak in the world is believed to be a new introduction to humans, except for the first one in Wuhan in late 2019.
<ul>
<li>This observation contrasts with previous viral outbreaks such as SARS-CoV-1, HIV, and MERS, where multiple spillover events were documented. Furthermore, when SARS-CoV-2 was transmitted to minks, multiple spillover events occurred from humans to minks, resulting in numerous mutations. Subsequently, the virus was transmitted back from minks to humans.</li>
</ul>
None of the millions of SARS2 sequences are phylogenetically distant from the Wuhan strains, suggesting that they are not a separate introduction event.
<ul>
<li>There is a claim that lineages A and B were introduced separately, but this is contested. Additionally, both lineages were present in Wuhan simultaneously.</li>
</ul>

No intermediate host found & No additional spillover

While RaTG13 was <a href="https://www.nature.com/articles/s41586-020-2951-z" target="_blank">fully sequenced in 2018</a> , a paper published by WIV researchers <a href="https://www.nature.com/articles/s41586-020-2012-7" target="_blank">in February 2020</a> implied that RaTG13 had been sequenced only after the COVID-19 outbreak (“We carried out full-length sequencing on this RNA sample”). However, the WIV paper referenced RaTG13 without mentioning that it was previously known and sequenced as BtCoV/4991.
<ul>
<li>Long after the relationship was pointed out, <a href="https://www.nature.com/articles/s41586-020-2951-z" target="_blank">an addendum published in November 2020</a> by Shi and her co-authors confirmed that BtCoV/4991 is indeed identical to RaTG13.</li>
</ul>
Later, when publishing the SARS2 RaTG13 alignment, they cut it right before the FCS, where the very obvious insert would appear

WIV confirmed RaTG13 as BtCoV/4991; important details omitted

The database of bat and rodent viruses with 80,000 samples went offline and then went back on for a short while in September 2019 and didn't come back online.
<ul>
<li>On December 30, when Dr. Shi Zheng-Li was informed of the COVID-19 outbreak, <a href="https://nypost.com/2020/05/06/what-is-china-covering-up-about-the-coronavirus-devine/" target="_blank">changes were made</a> to her bat virus database. She had previously said that it included data on virus variants in other wild animals, but certain terms such as “bat,” “rodent,” and “virus” were replaced with "wild animal samples,” “viral pathogen data,” “emerging infectious diseases,” and “cross-species infection.” The database was later deleted.</li>
</ul>
The WIV claimed a cyber-attack on the database.

Disruption of the bat and rodent viruses database

On December 10th, <a href="http://web.archive.org/web/20200414022725/https://twitter.com/VirusWhisperer/status/1204398902218297346" target="_blank">Shi Zhengli went to Singapore</a> for a conference about the Nipah virus.
On January 15th, a month after the outbreak started, the team <a href="https://archive.is/LfZN1" target="_blank">went out to dinner</a> in Wuhan.

WIV scientists acted normally at the beginning of the pandemic

Chinese officials <a href="https://www.businessinsider.com/china-confirms-that-it-destroyed-early-samples-of-new-coronavirus-2020-5?r=US&IR=T" target="_blank">confirmed </a> that they had destroyed samples of the novel coronavirus, claiming it was done "to prevent the risk to laboratory biological safety and prevent secondary disasters caused by unidentified pathogens."

China destroyed some samples, but said it was done for safety reasons.


At first, China <a href="https://www.bbc.com/news/world-asia-china-52420536" target="_blank">rejected the proposal</a> for an independent international investigation into the origins of the coronavirus, which was to be led by the WHO. They argued that such requests were politically motivated and asserted that focusing on an investigation would divert both attention and resources from their ongoing efforts to combat the pandemic.

China did not give the WHO direct access to labs

<a href="https://www.wsj.com/articles/who-china-hunt-covid-origins-11616004512" target="_blank">Strong limitations were imposed</a> on the WHO team that was eventually allowed in, a year after the initial outbreak.
<ul>
<li>China secured veto rights over participants and insisted its scope encompass other countries as well.</li>
<li>The team didn't receive raw data regarding the earliest infections.</li>
<li>WHO team members said they didn’t have the mandate, expertise, or access to investigate a potential lab leak.</li>
</ul>
<a href="https://globalnews.ca/news/7624297/china-coronavirus-probe-access-who/" target="_blank">Peter Daszak said</a> the WHO team had full access to everything they asked for.
<ul>
<li>However, he is motivated to focus on a zoonotic source, as he himself supported gain-of-function research, <a href="https://www.youtube.com/watch?v=IdYDL_RK--w&t=1793s" target="_blank">including coronavirus research</a> and <a href="https://www.nature.com/articles/d41586-020-02473-4" target="_blank">partnered with the WIV</a></li>
</ul>

Strong limitations were imposed on the WHO investigative team

The official Chinese response was not transparent: they did not share information and <a href="https://www.nytimes.com/2020/02/01/world/asia/china-coronavirus.html" target="_blank">silenced Chinese doctors</a> who were talking about COVID-19 and <a href="https://www.businessinsider.com/china-confirms-that-it-destroyed-early-samples-of-new-coronavirus-2020-5?r=US&IR=T" target="_blank">destroyed some samples</a> of SARS-CoV-2.
China <a href="https://www.businessinsider.com/china-confirms-that-it-destroyed-early-samples-of-new-coronavirus-2020-5?r=US&IR=T" target="_blank">destroyed some samples</a>, but said it was done for safety reasons.
China did not give the WHO direct access to labs.
<a href="https://www.wsj.com/articles/who-china-hunt-covid-origins-11616004512" target="_blank">Strong limitations were imposed </a>on the WHO team that was eventually allowed in, a year after the initial outbreak.

The official Chinese response was not transparent

We don't have any studies in which Chinese authorities tried to trace contact back, up to the earliest possible infection. This could have found patient zero, which could have led to the animal, and closed the case. This would also help in preventing future spillovers, not just in investigating the origin of COVID-19. This would be a relatively cheap effort mostly based on interviewing a few hundred patients and contacts, until possibly reaching patient zero. No such effort was published.

No published contact tracing efforts

The Chinese government <a href="https://thediplomat.com/2020/06/china-delayed-releasing-coronavirus-info-frustrating-who/" target="_blank">first tried to prevent publication of Covid-19's gene sequencing</a> , and only shared it after a private researcher had published the sequencing online.
On January 11, Shanghai’s Zhang finally published the coronavirus sequence ahead of health authorities on virological.org, used by researchers to swap tips on pathogens. It was only then that the Chinese CDC, Wuhan Institute of Virology and the Chinese Academy of Medical Sciences raced to publish their sequences, doing so on January 12.

Not sharing genomes

Up to this point, there has been no individual or entity associated with the WIV who has come forward to claim responsibility for any potential involvement or to blow the whistle on any misconduct linked to SARS Cov-2, not even through anonymous channels.

No whistleblower

The intelligence agencies have not provided any substantial information.
There were some claims of sick people at WIV (see below “No known WIV infections”) but government sources are not generally reliable.

No strong leaks / intelligence

A <a href="https://www.wsj.com/articles/intelligence-on-sick-staff-at-wuhan-lab-fuels-debate-on-covid-19-origin-11621796228" target="_blank">U.S. intelligence report claimed</a> that three researchers at the Wuhan lab fell ill in November 2019, but the exact illness is not known and the report cited anonymous government sources. Based on our experience in previous analyses, we give little weight to anonymous government sources.

US intel report claimed 3 WIV workers ill in Nov. 2019

However, the WIV claimed that researchers <a href="https://news.yahoo.com/team-scientist-wuhan-lab-workers-235458092.html?guccounter=1" target="_blank">tested negative</a> for SARS-CoV-2, in a self-reported study.
The staff health monitoring program <a href="https://www.who.int/publications/i/item/who-convened-global-study-of-origins-of-sars-cov-2-china-part" target="_blank">did not report COVID-19 or other respiratory illnesses</a> during the weeks or months prior to December 2019. They reported <a href="https://www.who.int/publications/i/item/who-convened-global-study-of-origins-of-sars-cov-2-china-part" target="_blank">no serological evidence of SARS-CoV-2 infection</a> in lab workers.
While there were no reported infections in the core research staff, there was no mention of testing for non-research staff at the facility.

The WIV claimed that researchers tested negative for SARS-CoV-2

It has been claimed that Huang Yan Ling, who worked at the WIV, could not be located after rumors that she was the first to contract COVID-19.
<ul>
<li>Ling <a href="https://www.dailytelegraph.com.au/lifestyle/health/coronavirus-nsw-dossier-lays-out-case-against-china-bat-virus-program/news-story/e1833c0893bfdac16d21dca6ce364bfb" target="_blank">disappeared</a> , and her biography and image were deleted from the WIV website.</li>
<li>Chinese officials <a href="https://www.nzherald.co.nz/world/covid-19-coronavirus-wuhan-virus-lab-scientist-thought-to-be-patient-zero-still-missing/VA4TIO726U7P36ZQG2WLHDDTNA/" target="_blank">censored online reports</a> about Ling.</li>
<li>An impersonal message <a href="https://www.nzherald.co.nz/world/covid-19-coronavirus-wuhan-virus-lab-scientist-thought-to-be-patient-zero-still-missing/VA4TIO726U7P36ZQG2WLHDDTNA/" target="_blank">was posted</a> from Ling’s WeChat account asking people to ignore rumors about her.</li>
<li>A Chinese news agency <a href="https://www.nzherald.co.nz/world/covid-19-coronavirus-wuhan-virus-lab-scientist-thought-to-be-patient-zero-still-missing/VA4TIO726U7P36ZQG2WLHDDTNA/" target="_blank">claimed that it had spoken</a> with her new employer but provided no other details.</li>
</ul>

Suspicious response to Huang Yan Ling’s disappearance 

We do not have any reliable information about infections in WIV lab workers.

No known WIV infections

The explanation of academic competition is unlikely in a major pandemic where we would expect other incentives to prevail.
Although the database's downtime in September may not be particularly relevant due to its early timing, it raises questions about why continuous, unrestricted access for virologists was not ensured. Such access could have had important public health implications and could have helped alleviate suspicions surrounding the WIV’s involvement. The claim of a cyber attack by the WIV appears questionable, given that alternative methods for database sharing could have been explored and implemented.
However, since some behavioral evidence supports each hypothesis, and the probability effect is relatively small, for simplicity we consider them to cancel each other out in terms of probability effects.

The best explanation for the pieces of behavioral evidence that are indicative of a lab leak is a mix of general mistrust and academic competition.
<ul>
<li>The suspicious behavior of WIV such as not sharing databases during a pandemic and acting strangely around RaTG13 is best explained by the context of international academic competition, which makes WIV unwilling to share their research at an early stage.</li>
<li>The best explanation for the behavior of the Chinese government is that they mistrust that an international investigation will be unbiased</li>
</ul>
The other behavioral evidence lean toward zoonosis and are therefore expected

General mistrust and academic competition

These explanations are likely. In addition, the Sverdlovsk anthrax leak, denied for years by the Soviet Union, provides a historical antecedent for the possibility of the absence of leaks or whistleblowers under such conditions.
However, since some of the evidence supports each hypothesis, and the probability effect is relatively small, for simplicity we consider them to cancel each other out in terms of probability effects

The best explanation for the pieces of behavioral evidence that are indicative of a zoonosis is that a small group of people knew about it and covered it up, combined with the fact that SARS Cov-2 has only mild symptoms in young people (to explain the lack of known WIV infections or the normal behavior of some WIV workers at the beginning of the pandemic).
<ul>
<li>The missing behavioral evidence under lab leak (absence of whistleblowers or leaks) are best explained by a cover-up by a small amount of people, especially given the risks involved. If the pandemic was a result of an incident at the WIV, most WIV personnel might not be aware of the initial leak, and those who did know might choose to hide it.</li>
<li>The lack of strong intelligence is likely due to WIV not being a military institution and therefore not an interesting target for intelligence agencies.</li>
<li>The light symptoms of SARS-CoV-2 among young people can explain the normal behavior of WIV workers at the beginning of the pandemic (where extra care and frenzy would be expected) as well as the absence of known WIV infections.</li>
</ul>
The other behavioral evidence lean toward lab leak and are therefore expected

Small group of people in the know covered it up

There is contrasting behavioral evidence: on the one hand, there is behavioral evidence pointing to zoonosis, such as:
<ul>
<li>The regular behavior of WIV workers at the beginning of the pandemic</li>
<li>Some missing expected evidence : the absence of leaks, the lack of strong intelligence, the lack of reliable information on infections among WIV laboratory workers</li>
</ul>
On the other hand, there is behavioral evidence suggesting lab leak:
<ul>
<li>WIV suspicious behavior: not sharing databases during a pandemic and their strange behavior around RaTG13.</li>
<li>The behavior of the Chinese government, which was not transparent, severely restricted the WHO team, and did not publish contact tracing efforts.</li>
</ul>