Thread by @kakape

A new preprint by @PeterDaszak, @nycbat and others attempts to show where the next coronavirus pandemic is most likely to begin and argues that there may be 400,000 hidden infections with SARSr-CoVs every year.

Story is here (thread to come):
science.org/content/articl…

@PeterDaszak @nycbat First of all:
Yes, that is a really big number. And yes there is HUGE uncertainty in that.
The confidence interval goes all the way down to a single case and all the way up to more than 35 million!
We’ll get to that.
But let’s take a quick look at what the researchers did.

@PeterDaszak @nycbat They created a detailed map of the habitats of 23 bat species known to harbor SARSr-CoVs, then overlaid it with data on where humans live to create a map showing where the risk of spillover is highest: southern China, Vietnam, Cambodia, and on Java and other islands in Indonesia

@PeterDaszak @nycbat (Quick note: SARSr-CoVs are “SARS-related coronaviruses”. That’s the virus species that SARS-CoV and SARS-CoV-2 belong to.)

@PeterDaszak @nycbat The maps show where people are most likely to be infected with a bat SARSr-CoV, says @PeterDaszak.
“Just living there means you’re exposed: People are sheltering in caves, they’re digging guano out of caves, they’re hunting and eating bats.”
(None of this is the bat’s fault!)

@PeterDaszak @nycbat Daszak says the map can help to target surveillance to detect the next outbreak, but also to target interventions to change behaviour in high-risk areas and reduce the risk of spillover.

Now, on to THAT number...

@PeterDaszak @nycbat How did they get to 400,000 infections a year?
People living in the bats’ range TIMES likelihood of human-bat contact TIMES likelihood such a contact leads to an infection that produces antibodies TIMES likelihood detectable antibodies are due to an infection the previous year

@PeterDaszak @nycbat Now, obviously: The estimate is only gonna be as good as the data that goes into it.
In this case a crucial piece are studies looking at what percentage of people carry antibodies for SARSr-CoVs and related viruses in a certain area (serology data).

@PeterDaszak @nycbat Vincent Munster told me that the serology data “is significantly less strong than the other well developed datasets they used in their modelling exercise”
That’s because :
A) there are very few studies looking at antibodies
B) these tests can easily yield a false positive result

@PeterDaszak @nycbat That’s why @DFisman calls the modelling “shaky”: “I think if the seroprevalence estimate is way off, the whole thing collapses.”
He also told me 400,000 did not ring true, because you would expect such regular spillovers to be recognized (as they are for rabies, say)

@PeterDaszak @nycbat @DFisman So take the exact number with a huge grain of salt. But one important point here is that the paper does show how to get to a number. With better data we should be able to get to a better number.

@PeterDaszak @nycbat @DFisman As @angie_rasmussen wrote to me: "I think this paper is important because, even if that number is way off, it provides a roadmap for thinking about how to quantify zoonotic spillover, a challenge that is absolutely crucial to overcome for pandemic preparedness."

@PeterDaszak @nycbat @DFisman @angie_rasmussen And THAT’S what really fascinates me:
How little we know about spillovers.
It’s reasonable to assume that there are many more infections than the ones we notice that lead to a large outbreak. But is that hundreds more, thousands more or hundreds of thousands more? We don’t know.

@PeterDaszak @nycbat @DFisman @angie_rasmussen Take HIV: Last time I checked, HIV was thought to have crossed over to humans at least 13 times:
HIV-1: 4 times (twice from chimpanzees, once from gorillas, fourth could be either).
HIV-2: 9 times from sooty mangabeys
Only one of these, HIV-1 M, caused the pandemic

@PeterDaszak @nycbat @DFisman @angie_rasmussen So at least 13 spillovers.
But these are just the ones that were successful enough to set up chains of infections that we can still detect today.
How many spillovers vanished without a single onward transmission? Or after a few transmissions?

@PeterDaszak @nycbat @DFisman @angie_rasmussen Think about it this way:
Do bat viruses rarely infect humans and when they do they have a good chance of getting transmitted and causing a large outbreak?
Or do they infect humans all the time but mostly lack the right genetics to spread further?
That makes a difference!

@PeterDaszak @nycbat @DFisman @angie_rasmussen Rasmussen does not think the number 400,000 is outlandish given that many viruses might not infect enough cells (or the right cells) to be transmitted to another person, or might not be able to escape humans’ immune defenses or might not spread for a host of other reasons.

@PeterDaszak @nycbat @DFisman @angie_rasmussen She wrote: "400,000 infections/year in a region with likely hundreds of millions of bats and nearly half a billion people isn’t that many, even if the overall estimate misses the mark by a lot, if you consider the vast majority of these infections are asymptomatic dead ends"

@PeterDaszak @nycbat @DFisman @angie_rasmussen The fact that scientists like @angie_rasmussen and @DFisman can have such different takes on this number is another sign to me how little we really know.
And one of the perks of being a science journalist is that I can watch scientists as they try to learn more….