Thread by @wormmaps

Let's say that after COVID-19, you wanted to discover every single animal virus that can make a person sick, from Aichi to Zika.

How many are there?
How long would it take?
A little thread for a quiet night.

When you walk into a party, you meet the most strangers early in the night; the longer you’re there, the fewer new people you’re likely to meet, and the more repeats you hit.

Ecologists usually measure diversity the same way, using what they call “rarefaction curves.”

Every year, about 2-4 new animal viruses that can infect humans – zoonotic viruses – are discovered. Even though we’re being hit by more virus outbreaks every year, we started hit a plateau in zoonotic virus discovery over the 20th century.

In 2013, one study used rarefaction methods and guessed that there should be about 187 zoonotic RNA viruses, based on the 158 RNA viruses of humans they knew about – so about 10-20% still undiscovered.

futuremedicine.com/doi/pdfplus/10…

If you corrected for gaps where we might not have checked some virus families, you could maybe push that number up to ~300 zoonotic RNA viruses.

That'd bring you back down to something that makes sense: about half known to science, half undiscovered.

The same year, the PREDICT consortium’s team came up with a way to try to estimate global viral diversity!

mbio.asm.org/content/4/5/e0…

It had two steps:

1. Fully sequence one animal species: a flying fox bat. Use rarefaction, and estimate how many virus species there were total (~ 60).

2. Extrapolate out to all the mammal species on Earth.

Total: 320,000 mammal viruses.

In 2018, when the Global Virome Project was proposed, they added a second species – a monkey, with a lot of viruses – to that estimate.

It pushed the total way higher, to 1.67 million viruses in birds and mammals.

(You might’ve heard that number a lot lately.)

Of those viruses, they thought about 40% could infect humans, based on the percent of all the viruses we’ve discovered so far that are zoonotic. (Yes, sampling bias.)

That puts you in the 600,000-800,000 range for zoonotic viruses. Way more than 297.

Last year, we wrote our own paper about this!

nature.com/articles/s4155…

It turns out sampling hosts works a lot like rarefaction – the more hosts you sample, the fewer new viruses you find.

That’s actually true of any symbiosis – animals and viruses, anemones and clownfish, flowers and butterflies.

And that, so to speak, flattens the curve – pushing total mammal viruses down to about 50,000, and zoonoses to about 10,000.

If you really push, you can get about 37,000 DNA (5,000 zoonoses) and 18,000 RNA (7,500 zoonoses). Let’s call that 13,000 zoonotic viruses in mammals.

So now we have three estimates:
❓~300 zoonotic RNA viruses
❓13,000 zoonotic viruses (7,500 RNA)
❓800,000 zoonotic viruses

All produced by very smart scientists, who study emerging viruses for a living. What do we do with that? How do we make sense of it?

If I’m being honest: we haven’t yet, and it’s fair if you can’t. Last year, one of my strangest conversations about our paper included a colleague saying, “I mean it's a nice paper but, you know how it'll be cited: There’s 800,000 zoonotic viruses (but see Carlson 2019)”

But I think you deserve to know that the 1.7 million viruses number isn’t the only one out there, because it has some pretty big implications for how we move on after COVID-19.

Let’s talk about virus hunters.

The PREDICT program was funded for $207 million over 10 years.

PREDICT usually says they discovered 1,000+ viruses.

Fact check: they discovered 949 but that “discovery” seems to include 217 that have been previously known, like Ebola.

Of ~1,000 viruses, a lot of them were probably zoonoses, because smart people were looking in the right places.

But only 1 zoonotic virus discovery is known for sure (Bas-Congo virus). It's hard to know if a virus affects us just from a genome sequence.

thelancet.com/journals/lanmi…

Now the Global Virome Project has pitched that over a decade (just like PREDICT), they could discover 1.67 million viruses with ~$1.2 billion.

Techology is constantly getting better and cheaper, so they guessed it would cost a bit under $1,000 to discover each new virus.

But at PREDICT’s rate - a global program of brilliant scientists that ended last year, using modern tech - they’d only discover 4,243 viruses.

Now that 4,243 might be ~10% of the global virome if there’s 50,000 viruses;

It would be < 1% if there's 1.7 million viruses total.

All of this matters because we - scientists - are asking for your help to prevent the next pandemic.

The 1.7 million viruses number is going around in press, because we want to keep funding important scientific work in the United States and abroad, with your taxpayer money.

After COVID-19, we need a Global Virome Project.

But what GVP do we need?

To me, the most exciting ideas pitched by GVP are things like: developing a universal betacoronavirus vaccine (proposed before COVID-19!).

Ones that don’t count on virus hunting & fuzzy math 🥴