4 ms·

Has anyone analyzed this with a Bayesian statistical approach? Considering that a zoonotic transfer could happen in a very large number of places on earth, I wo

Has anyone analyzed this with a Bayesian statistical approach? Considering that a zoonotic transfer could happen in a very large number of places on earth, I would imagine that the odds of an outbreak in any particular city would be quite small? With that in mind, how many virology labs are there doing gain-of-function research on coronaviruses? How many have had safety issues in the past? How many times have pathogens escaped? Also, how many zoonotic transfers have there been where no infected specimens of the host species could be found after a concerted investigation? It seems like one should be able to come up with reasonable estimates for most of these numbers with some digging. I'm not qualified to make conclusions one way or the other, but it feels to me like this calculation would weight heavily toward a lab leak. To a layman's eyes, it feels a bit like we've got a lung cancer patient who just so happened to be a pack-a-day smoker, and everyone's searching furiously for any explanation that isn't the cigarettes.

rllearneratwork 5y ago
>>"Has anyone analyzed this with a Bayesian statistical approach?" - YES. And they give lab leak theory over 80% likelihood https://www.rootclaim.com/analysis/What-is-the-source-of-COVID-19-SARS-CoV-2 https://www.rootclaim.com/analysis/What-is-the-source-of-COV...
- mhh__ 5y ago
  How can you assign a probability/likelihood to vague statements like "There is some weak evidence regarding lax security and procedures at the Wuhan Institute of Virology"
  - fastball 5y ago
    Probabilities determined from Bayesian analyses like this one aren't really intended to be perfect – it is more of an exercise to help you think through the various different possibilities in a more structured way. In other words, Bayesian estimates should always be taken as relative to all other related estimated outcomes rather than as absolutes in isolation (even thought you are arriving at a specific "probability").
    - mhh__ 5y ago
      I am aware. This is why I find it particularly grating if you even pretend it's not subjective.
orlovs 5y ago
U know, this theory was dead due fact controversial politician said it loud. Ofc he had to keep mouth shut until point there were hard evidence. Still one of main reasons why it was dismissed who said, not what being said. As global pandemic is complex issue where are many bolts and knobs. However, it still would be usefull why engine blowed up, not just why pilots acted badly during crash
handmodel 5y ago
This is my thought. I think it is hard to do a Bayesian analysis here and come up with priors everyone can agree on but it always seemed to me it would be like if a new Mad Cow disease started near Cambridge, Mass. If a new cow disease arose naturally it would probably be in a cattle farm with tons of animals being slaughtered - not next to a lab that studies prions.
- actually_a_dog 5y ago
  The problem with that is there are way more pathogenic viruses originating in animals than prion diseases, and (probably?) way more virology labs than labs studying prion diseases. That makes both factors in the equation much larger in the case of viruses than prion diseases. Just off the cuff, I'd guess that would make the overall probability of a zoonotic virus emerging near a city with a virology lab at least an order of magnitude greater than your prion disease scenario.
thurn 5y ago
I assume virology labs are not randomly distributed across the country, e.g. there could be many reasons why a lab studying coronaviruses might be located in close proximity to an area with a high risk of outbreak.
- tripletao 5y ago
  The closest animal virus to SARS-CoV-2 was found in nature about 900 miles from Wuhan (RaTG13, in Mojiang). Dr. Shi herself thinks spillover in Hubei is unlikely: > We have done bat virus surveillance in Hubei Province for many years, but have not found that bats in Wuhan or even the wider Hubei Province carry any coronaviruses that are closely related to SARS-CoV-2. I don't think the spillover from bats to humans occurred in Wuhan or in Hubei Province. https://www.sciencemag.org/sites/default/files/Shi%20Zhengli%20Q%26A.pdf https://www.sciencemag.org/sites/default/files/Shi%20Zhengli... The WIV's program to sample novel SARS-like viruses from nature and manipulate them in the lab was the biggest in the world. The only other group that comes close is Ralph Baric's, and Baric has joined the letter in Science calling for further investigation of the origin: https://science.sciencemag.org/content/372/6543/694.1 https://science.sciencemag.org/content/372/6543/694.1

rllearneratwork 5y ago

>>"Has anyone analyzed this with a Bayesian statistical approach?" - YES. And they give lab leak theory over 80% likelihood https://www.rootclaim.com/analysis/What-is-the-source-of-COVID-19-SARS-CoV-2 https://www.rootclaim.com/analysis/What-is-the-source-of-COV...

mhh__ 5y ago
How can you assign a probability/likelihood to vague statements like "There is some weak evidence regarding lax security and procedures at the Wuhan Institute of Virology"
- fastball 5y ago
  Probabilities determined from Bayesian analyses like this one aren't really intended to be perfect – it is more of an exercise to help you think through the various different possibilities in a more structured way. In other words, Bayesian estimates should always be taken as relative to all other related estimated outcomes rather than as absolutes in isolation (even thought you are arriving at a specific "probability").
  - mhh__ 5y ago
    I am aware. This is why I find it particularly grating if you even pretend it's not subjective.

orlovs 5y ago

U know, this theory was dead due fact controversial politician said it loud. Ofc he had to keep mouth shut until point there were hard evidence. Still one of main reasons why it was dismissed who said, not what being said. As global pandemic is complex issue where are many bolts and knobs. However, it still would be usefull why engine blowed up, not just why pilots acted badly during crash

handmodel 5y ago

This is my thought. I think it is hard to do a Bayesian analysis here and come up with priors everyone can agree on but it always seemed to me it would be like if a new Mad Cow disease started near Cambridge, Mass. If a new cow disease arose naturally it would probably be in a cattle farm with tons of animals being slaughtered - not next to a lab that studies prions.

actually_a_dog 5y ago
The problem with that is there are way more pathogenic viruses originating in animals than prion diseases, and (probably?) way more virology labs than labs studying prion diseases. That makes both factors in the equation much larger in the case of viruses than prion diseases. Just off the cuff, I'd guess that would make the overall probability of a zoonotic virus emerging near a city with a virology lab at least an order of magnitude greater than your prion disease scenario.

thurn 5y ago

I assume virology labs are not randomly distributed across the country, e.g. there could be many reasons why a lab studying coronaviruses might be located in close proximity to an area with a high risk of outbreak.

tripletao 5y ago
The closest animal virus to SARS-CoV-2 was found in nature about 900 miles from Wuhan (RaTG13, in Mojiang). Dr. Shi herself thinks spillover in Hubei is unlikely: > We have done bat virus surveillance in Hubei Province for many years, but have not found that bats in Wuhan or even the wider Hubei Province carry any coronaviruses that are closely related to SARS-CoV-2. I don't think the spillover from bats to humans occurred in Wuhan or in Hubei Province. https://www.sciencemag.org/sites/default/files/Shi%20Zhengli%20Q%26A.pdf https://www.sciencemag.org/sites/default/files/Shi%20Zhengli... The WIV's program to sample novel SARS-like viruses from nature and manipulate them in the lab was the biggest in the world. The only other group that comes close is Ralph Baric's, and Baric has joined the letter in Science calling for further investigation of the origin: https://science.sciencemag.org/content/372/6543/694.1 https://science.sciencemag.org/content/372/6543/694.1