Comment by anikan_vader | Hacker News Reader

Comment by anikan_vader | Hacker News Reader

Parent Full thread

It is interesting to think about the origins of Covid-19 with Bayesian statistics in mind. We believe that the virus originated in Wuhan, and the two competing "mainstream" theories are whether the virus came from a wet market or a bio lab.

I don't know exactly how many wet markets exist in China, but I would guess that there are thousands spread across China. A wet market is simply a place where fresh produce/meat is sold (think farmer's market). "In Hong Kong, for example, there is a widespread network of wet markets where thousands of locals shop everyday for their meat and vegetables. There is one in almost every district and none of them trade in exotic or wild animals." [1] Of these $\Theta(1000)$ wet markets, probably a few are located in Wuhan.

Notably, there is only one level 4 bio lab in China, and it happens to be located in Wuhan. [2]

You can't conclude anything with absolute certainty doing this kind of probabilistic analysis, but this line of argument swayed me to seriously consider the possibility that the virus originated in the lab.

And, of course, note that we do not have to assume any malicious intentions. It's easy to imagine a virus accidentally escaping the lab, regardless of whether it evolved naturally or with some human intervention to facilitate research. A State Department cable from 2018 wrote that the lab "has a serious shortage of appropriately trained technicians and investigators needed to safely operate this high-containment laboratory.” [3]

[1] https://www.cnn.com/2020/04/14/asia/china-wet-market-coronav...

[2] https://www.livescience.com/china-lab-meets-biosafety-levels...

[3] https://www.washingtonpost.com/national-security/state-depar...

Footnote: Yes, I know that $\Theta(1000)$ just means "some constant" from a mathematical perspective. But I think the idea is clear, even if the mathematical statement taken at face value conveys no information.

That's an interesting argument. I guess one critical piece of data would be - is the probability of Covid-19 originating from wet markets the same across all of China?

If every wet market had the same probability, then yeah, it's odd that the alleged wet market is one of thousands, but just happened to be next to the bio lab in Wuhan.

But if the probability was unevenly distributed and such a virus originating was highly likely in only a few dozen wet markets, Wuhan included, then the probability of it being a coincidence is much higher.

Interesting. But also to GP’s point, I never thought of that perspective.

It is quite unsettling to realise the strangeness of odds that it occurred right next to the Wuhan lab...

"Right next to" being 14 km away, on the other side of a major metropolis.

Nobody knows where the outbreak actually began, though. Wuhan was just the first place it was detected. It could easily have begun somewhere in the countryside, only taking off when an infected person went to a major city. Just look how long it circulated undetected in the US - for more than a month - before anyone noticed.

The wet market and lab are not the only two options. The millions of people who come into contact with animals infected by novel coronaviruses on farms or in nature every year are, by far, the most likely origin of the outbreak.

One the one hand, you have a handful of highly trained technicians basically wearing spacesuits, and undergoing strict decontamination procedures.

On the other hand, you have millions of people in the countryside who live around bats (a few percent of which are infected with SARS-like coronaviruses at any given time), or who live around livestock that are potentially exposed to bats. You have people who go into bat caves specifically to collect guano (i.e., bat poop). And there's evidence that a few percent of people in villages near these bat populations have actually been infected with novel bat coronaviruses (based on serology).

While you worry about a few people in the lab, there are millions of people in far more risky everyday situations you're ignoring.

Interesting thought process.

One minor correction: there's two BSL-4 labs in China. The other is in Harbin, Heilongjiang.

https://en.wikipedia.org/wiki/Biosafety_level#List_of_BSL-4_...

The issue is that the lab was placed in Wuhan to begin with because it was likely a coronavirus could spillover. That throws a wrench into that kind of analysis.

Comparing wet market vs. lab escape based on numbers of each isn't that useful to analyze.

For one thing, those are far from the only places the virus could have jumped to humans. (In fact, investigations of early cases in Wuhan show that while the market appears to be the site of a super spreader event, it wasn't the origin of the spread of the virus in humans.)

Also, by comparing counts of wet markets and bio labs, you ignore that the likelihood of spread from each are not comparable -- they'd have to have something like the same order of magnitude of likelihood for your analysis to have meaning. Obviously, that's rather unlikely to be true.

> You can't conclude anything with absolute certainty doing this kind of probabilistic analysis

I don't think you can conclude anything from your analysis considering all the holes in it.

Additional data point: Of the earliest identified cases, less than half had any connection to the wet market.

Also there's one misconception here:

> A wet market is simply a place where fresh produce/meat is sold (think farmer's market).

Farmer's markets, if they even have live animals, are kept distanced from food being sold. Wet markets, live animals are often being sold for food and/or butchered in front of you upon sale.

> "none of them [wet markets] trade in exotic or wild animals" That's not what we've been hearing at all. Did the Huanan Seafood Market have live wild animals or not? If not, why does your quote contradict literally everything I've ever read about this market?

Reply on news.ycombinator.com