Thread by @__ice9

For reference, here is the recent report alleging a potentially artificial origin for SARS-CoV-2, from virologist Li-Meng Yan.

zenodo.org/record/4028830…

en.wikipedia.org/wiki/Li-Meng_Y…

Yan recently fled Hong Kong and remains in hiding in the United States.

Key arguments arise from the rarity (in bat CoVs) of certain features allowing human transmissibility, poor quality data on closest purported relative RaTG13, and extremely close similarities (unique vs. all other sequences) to strains ZC45/ZXC21 reported by military facilities.

Also noted: inability of RaTG13 to actually bind horseshoe bat ACE2, the suppression (by immediate physical lab closure) of a Chinese research group describing the closer relationship to ZC45/ZXC21 in a Nature article, and the fact that the only major differences are in Spike.

Overall, I find these claims more compelling than others to date.

The extreme and unique similarity to ZC45/ZXC21, apart from the only region needing significant changes in order to infect humans, strongly suggests gain-of-function research.

The authors describe a plausible method for constructing SARS-CoV-2 from ZC45/ZXC21, given a few months of cell culture work.

Natural recombination appears less plausible, given such targeted deviations from the ZC45/ZXC21 strain. Other proteins should have also been altered.

Strikingly, SARS-CoV-2 Spike RBM instead resembles SARS-CoV RBM (rather than ZC45/ZXC21 RBM), but binds human ACE2 with much greater affinity.

Meanwhile, the furin site of SARS-CoV-2 is completely unique relative to all other known lineage B betacoronaviruses.

More on ZC45/ZXC21:

ncbi.nlm.nih.gov/pmc/articles/P…

advances.sciencemag.org/content/6/27/e…

Natural recombination is not necessarily ruled out, in principle.

Certain reported pangolin CoV RBDs also resemble SARS-CoV-2 RBD, and another alleged close relative, RmYN02, has also been recently reported.

However, the pangolin theory has its own discrepancies, identified by researchers at the Broad Institute of MIT and Harvard.

news-medical.net/news/20200708/…

Essentially, only one sample matches.

The group contends that the pangolin in question was more likely infected by humans.

Meanwhile, RmYN02 (also from Yunnan, like SARS-CoV, the type most studied in Wuhan) again does not match SARS-CoV-2 Spike, yet has strong similarities in nearly all other proteins.

cell.com/current-biolog…

It is also not clearly closer than ZC45/ZXC21, by the author's own data.

So then:
- RaTG13 is allegedly 'lost,' still lacks furin site etc.
- RmYN02 is no closer than ZC45/ZXC21, studied by same groups regardless
- pangolins suspected mere recipients, not intermediate hosts
- all CoVs related to ZC45/ZXC21 lack SARS-CoV-2's human adaptations

Then:

Whence comes Spike RBD?

Why are the only major differences in Spike?

And why was ground zero in Wuhan, only a few kilometers from arguably the world's foremost laboratory focused on these viruses, instead of in Yunnan or Zhoushan in the first place?

The Wuhan Institute of Virology maintains multiple facilities-- one centrally in Wuchang District, and one at Huanan New Materials Industrial Park in Jiangxia District.

The Wuchang facility is directly across the Yangtze River from the Huanan Seafood Market in Jianghan District.

Regardless, the seafood market is not believed to be the original site at which the virus first entered the human population.

livescience.com/covid-19-did-n…

Rather, it appears to have been merely the first known location of a human superspreader event.

What is the probability that the infected person buying groceries there worked at WIV?

It is conveniently located near the lab and residential districts.

1-2 week latency period, pre-symptomatic transmission.

The list of researchers infected in BSL3 and BSL4 labs is very long.

In general, I am mainly interested in clinical interventions to reduce the severity and transmissibility of COVID-19.

Origin theories are of somewhat secondary importance.

But overall, this investigation changed my view.

The simplest theory is that SARS-CoV-2 arose in bat CoV gain-of-function research at WIV, exiting via an infected human.

It explains location, timing, deviation from reported sequences, and oddly combined human-relevant features.

SARS-CoV-2, clinically and genetically, shows no indication as any kind of 'bioweapon' or military technology.

It is not especially virulent. SARS-CoV is far deadlier.

No new virulence factors or even new genes appear. No recombination from other types of viruses occurs.

It was released on home territory, seemingly by accident, near a CAS lab rather than a military one--

A CAS lab conducting gain-of-function research on CoVs very closely related to it.

1000 km from Zhoushan.

1900 km from Yunnan.

There are no bats in Wuhan.

Instead, it resembles exactly what one would expect for gain-of-function research.

Its skeleton closely resembles an earlier-reported strain.

The only appreciable differences against any known close relatives are in precisely the locations needed to gain human transmissibility.

So, someone did their job.

They proved, beyond any shadow of a doubt, that it is possible for a SARS-like coronavirus to gain the ability to infect and widely transmit among humans.

This was always a major goal in the field-- to understand if, and how.

thelancet.com/journals/lanin…

Such research has remained controversial.

It is not clear how well it parallels natural viral evolution.

For expediency, large selective pressures are applied at particular sites in cell culture. Limited synthetic methods may occasionally be used as a shortcut.

Fouchier's landmark paper...

ncbi.nlm.nih.gov/pmc/articles/P…

freethink.com/articles/gain-…

... provoked extensive objections, from basic plausibility to actionability to the risk of artificial pandemics.

cidrap.umn.edu/news-perspecti…

science.sciencemag.org/content/342/61…

europepmc.org/article/pmc/pm…
.

>"I was intrigued," says Ron Fouchier, in his rich, Dutch-accented English, "in how little things could kill large animals and humans."

>It's late evening in Rotterdam as darkness slowly drapes our Skype conversation.

>28 Mar, 2020 09:00 AM

Be that as it may, the techniques do leave a certain kind of 'fingerprint,' distinct both from natural evolution and from outright biological weapons programs.

1) start with a natural strain
2) optionally, hybridize with another strain, simulating co-infection
3) passage, optionally with mutagens, to acquire further specific desired features
4) optionally, edit in small tweaks that did not arise, hand-wave result as plausible in paper

5) make up something about how this provides a valuable early warning to prepare for a possible pandemic, shows that such-and-such virus could theoretically gain terrible new capabilities outside of your ever-so-responsible lab, etc.
6) angle for more grant funding.

But of course, the real reason:

>"I was intrigued," says Ron Fouchier, "in how little things could kill large animals and humans."

'I like playing with invisible transmissible death.'

Gain-of-function research should be banned by international law. It has only been 8 years.

One important point, not addressed in the Yan et al. manuscript--

SARS-CoV-2 still has dN/dS values (including in Spike) vs. related CoVs that are low and in typical ranges for RNA viruses.

This argues against extensive positive selection.

https://twitter.com/__ice9/status/1306049991652278272?s=20

Hence, if the broader claim of lab escape is valid, then it was likely either a mostly natural strain, or one modified chiefly by recombination against one or more other natural strains.

Not all samples sequenced are disclosed, making it difficult to exclude either possibility.

As Twitter has apparently suspended the account of the author, and I am not particularly interested in censorship as a response to controversy, my account will be set protected for a bit.

https://twitter.com/kacper_fronc/status/1306067175493398528?s=20

Further intrigues--

This grant proposal is nearly a verbatim description of a program of optimizing southern Chinese bat CoVs by ACE2-binding-based methods to gauge the risk of human transmission from nature.

https://twitter.com/__ice9/status/1306100230773968897

The project scope would have involved large numbers of strains closely related to SARS-CoV-2, handled systematically, likely without individual sequence disclosures-- and the dates align as well, mid-2019 to mid-2020 anticipated work.

Further discussion:
independentsciencenews.org/commentaries/a…

Article discussing the collapse of the pangolin papers, mainly focusing on work by @Ayjchan

Also discusses some of the same unusual attributes of SARS-CoV-2 mentioned in the Yan paper cited at the top of this thread.

Consistent with the grant language.

https://twitter.com/talentpharmNYC/status/1304131865423622146?s=19

Article on synthetic virology--

Example reconstruction of infectious SARS-CoV-2 virions from sequence data within about one week of work.

https://twitter.com/K_G_Andersen/status/1306037078602383360?s=20

Of course, this one does not prove anything about the origin of any given virus. Merely an available technology.

In attempting to confirm my procedural impressions of the grant work above, from descriptions by others closer to the subject matter, I encountered a very close match to the nature of the work I had suggested was involved. This strengthens the claim.

https://twitter.com/Ayjchan/status/1297150935744090112?s=20

... they just came right out and said it.

https://twitter.com/OneOfMany3421/status/1306671434790510595?s=19

Thread:

https://twitter.com/PeterDaszak/status/1197631383470034951?s=19

Stamps:

https://twitter.com/tweet_stamp/status/1306673528121503744?s=19

https://twitter.com/tweet_stamp/status/1306673603736416258?s=19

Further discussion here also:

https://twitter.com/R_H_Ebright/status/1253132609238024195?s=19

Systematic gain-of-function studies against hACE2 and humanized model animals using recombinant methods to modify Spike across dozens of novel undisclosed bat CoVs had recently begun at WIV at the time of the outbreak.

I have also gone to the trouble of submitting each pdf in the thread at their source to the Internet Archive.

web.archive.org/web/2020091719…

web.archive.org/web/2020091719…

web.archive.org/web/2020091719…

Published and available elsewhere, of course, but hopefully a useful combined collection.

Minor note--

The gain-of-function studies disclosed in papers and grants associated with the WIV's bat CoV research do not imply the design of a particular single 'candidate' strain.

Past work developed several, reporting on various characteristics.

https://twitter.com/__ice9/status/1306694599453429761?s=19

Meanwhile, evidence from the RaTG13 sequencing datasets strongly implies that southern Chinese bat CoV full genome sequencing had been ongoing without disclosure of the results to international databases for at least two years prior to the outbreak.

https://twitter.com/Ayjchan/status/1306993218408587267?s=19

Also, direct evidence that the backbone in these studies was not always a pseudovirus or SARS-CoV, even in 2008.

This paper did the reverse: generate synthetic consensus Bat-SCoV, swap in SARS-CoV RBD, test in cells and mice.

pnas.org/content/105/50…

Grant merely generalized.

Likewise, the 2015 Baric and Shi paper (includes researchers at WIV by 2019) used mouse-adapted SARS-CoV as a backbone instead of ordinary SARS-CoV.

med.unc.edu/orfeome/files/…

Fully synthetic recombinant virus used for an emergence study. The grant extended this specific work.

The grant described screening a new sample of apparently >50 SARS-like CoVs vs. human ACE2 and testing recombinants in humanized mice.

Note emphasis on pathogenicity in all of these papers. The backbone matters (e.g. SARS is deadlier in part because of better IFN suppression).

Hence, to properly understand potential emergence harms from this broader sampled set in humanized mice, the most logical approach is using both backbones and RBDs drawn from the set.

Fully synthetic methods and whole-genome sequences were already available, as shown above.

The goal? Likely to assess human emergence risk from the Mojiang Yunnan cave. 6 miners were infected; 2 died.

Shi lied, claiming fungal infection:
scientificamerican.com/article/how-ch…

But a translated thesis proves it was a bat CoV:
independentsciencenews.org/commentaries/a…

This cave also yielded: RaTG13.

In spite of initial doubts, the latest analysis linked above via @Ayjchan strongly suggests RaTG13 is indeed real.

And RaTG13 is the closest known match to SARS-CoV-2, >96%.

So the "bat woman of China," who contributed to the hunt for SARS and this covered-up CoV of the Mojiang miners ('threat for human emergence' it already happened) had a grant for systematic recombination studies-- including further Yunnan samples.

Wuhan is 1900 km from Yunnan.

96% similarity between RaTG13 and SARS-CoV-2, grant funding for recombinantly optimizing hACE2 binding over this class of CoVs and infecting humanized mice, undisclosed existence of the full RaTG13 sequence and likely others since at least 2017, 4 separate cave sampling trips,

... multiple trivially demonstrable lies from Shi, the appointment of Daszak himself to head the Lancet commission on the origins of COVID-19 (he likely *is* the origin of COVID-19), immediate statements against lab involvement...

covid19commission.org/peter-daszak

telegraph.co.uk/global-health/…

... RaTG13 and other bat CoVs from the same deadly Yunnan cave fitting the stated inclusion rule for use in the hACE2-optimized recombination grant, SARS-CoV-2 indeed binding human ACE2 hundreds of times better than SARS-CoV, winter in a city outside known ranges of such bats,

... fully synthetic recombination techniques and full genome data already available, prior use of non-SARS-CoV backbones and a clear justification for doing so again, a demonstrated history of refusing to disclose sequences and lying about both culture work and deadly pathogens,

... ground zero a human superspreader event at a market down the highway from the WIV CAS office and a few hundred meters from the Wuhan CDCP, *right around a biosafety conference with staff from the BSL3/4 lab* known to be infecting animals at the time.

https://twitter.com/__ice9/status/1306984585096503297?s=19

One was likely sick and simply didn't know it yet, and that is all that would have been required.

The furin site facilitates pre-symptomatic spread, contrary to SARS-CoV which lacked one. It confused the WHO and national authorities. The lab likely didn't quarantine for it.