Thread by @MichaelWorobey

1: A thread connecting the dots between:

(1) @PeterDaszak et al's fascinating recent preprint on the *many* SARS-related CoV infections in humans per year in Southeast Asia, and

(2) The furin cleavage site of SARS-CoV-2, and

(3) Why Wuhan?

medrxiv.org/content/10.110…

2: The study uses a clever combination of data streams to estimate that 400,000 people per year are being infected by SARS-related coronaviruses.

3: The authors note that not all of those infections are likely to be transmissible human to human

I strongly agree with this. We would see multiple new pandemic origins every year if even 0.1% of these were viable human to human pathogens.

4: I think of pandemic origins as a layer of selective sieves, each with a finer mesh than the previous one.

The first sieve is lack of ecological opportunity. Most SARSr CoVs never have a chance to get from a bat or intermediate species into a human. But some do.

4: They are the ones that circulate in animals that get close to humans. Bats in caves near human settlements, or viruses in intermediate hosts that get to wet markets, for example, as with SARS1.

5: These lucky viruses encounter sieve 2: human defenses. Our bodies have wonderful 'restriction factors' that can prevent a virus from completing its life cycle.
One, for example, mutates the hell out of the genomes of baby viruses, preventing them from having their own babies.

6: But some have the right genome to sidestep these defenses and flourish within a human. These encounter sieve 3: inability to get to the next host.

We see this with avian flu. Avian H5N1s might be great at spreading deep in the lungs, killing their human host...

7: ...but don't do well in the upper respiratory tract, so they never move on to human #2.

A subset of SARSr-CoVs can get through this sieve, and can infect at some secondary human hosts.

8: But only a subset of that subset happen the have the right genetic stuff to get through sieve 4: inefficient human to human transmission.

SARS-CoV-2 has the right genetic stuff.

9: But even unicorns like SARS-CoV2 hit one last sieve (#5): not enough fertile soil. Joel Wertheim, @pekar @niemasd Konrad Scheffler and I showed that even a virus as contagious as SC2 would die out 70% of the time when dropped into a city like Wuhan...

science.org/doi/full/10.11…

10: ...and would die out 94.5 to 99.6% of the time it landed in a rural environment.

So, assuming 400,000 SARSr-CoV infections per year, and a new outbreak/pandemic every 20 years, SARS-CoV-2 is a 'one-in-8-million' virus, *not* just a random sample of bat SARSr-CoV diversity.

11: As such, we should not be surprised *at all* that SARS-CoV-2 has a rare set of genetic features that allow it to infect humans and transmit between them.

12: Its furin cleavage site is one such a feature.

It may or may not be vanishingly rare in nature - you might have to screen a million bat samples before you see it. But it's exactly the sort thing you expect in a unicorn that gets through sieves 1-4.

13: And a city like Wuhan is what you expect after sieve 5. Large, densely populated, served by multiple wet markets selling the same species involved in SARS1 emergence.

14: This is (part of the reason) why, even though I keep an open mind to lab leak scenarios, I think a zoonotic origin from a wet market, or markets, is so, *so* much more likely.