The origin of the coronavirus

August 17, 2021 • 1:00 pm

A new paper in Science (click on screenshot below, pdf here, reference at bottom) suggests that Covid-19—referred to in this study as SARS-CoV-2—likely originated in horseshoe bats that were collected in Yunnan in southern China, were not contaminations from the Wuhan Institute of Viriology (WIV), but were transferred to humans via an intermediate animal vector (probably a civet cat) in a wet market. Note that two of the authors are Chinese, and one might think that they have an interest in exculpating the WIV, but that opens up a whole can of worms that I’d prefer to avoid.


Here’s a phylogeny (family tree) of the “sarbecoviruses” that are evolutionarily closest to the Covid-19 virus, with the caption from the Science paper. Click on photo twice to make it really big.

Note that the three viruses closet to human coronavirus in sequence are all from areas close together in Yunnan, and all in species of horseshoe bats ( genus Rhinolophus, variants RpYN06, RmYN02, and PRC31). Viruses in pangolins (Manis javanica), also presumably derived from bats, are much less closely related, and thus unlikely as a source of human infection. The virus RaTG13, sequenced and kept at WIV, seems too distant from the human coronavirus to have been the source, and horseshoe bats are found not just in Yunnan, but are widely dispersed throughout China.

The authors posit that, since bats were not sold in Wuhan markets, another animal—they think the civet cat—is the likely transmitter of the virus to humans in a wet market, and this happened in about December of 2019. The bat virus may have gotten into a civet (or a raccoon dog, or a fox, or a mink) on one of the many farms where these animals are raised for sale as meat, and then transported to wet markets in other places in China.

As an interesting sidelight, the authors suggest that the spread of the coronavirus was promoted by a shortage of pork in China in 2019, which itself was due to swine flu that led to 150 million pigs being killed. They posit that other animals, like civet, could have replaced pork in the diet, and those animals would be intermediate vectors that led to the interspecies leap in late 2019. (Our own species is now considered the main vector for Covid-19!).  They suggest, alternatively, that the virus could have survived in frozen wild meat rather than in live animals sold in wet markets.

Finally, now that we’re the main host of the virus, the authors worry that we ourselves could infect other wildlife, which would then become reservoirs for evolution and re-infection (this is called “reverse zoonoisis”)

There are several questions that are unanswered in this short paper, but may be common knowledge. How do the authors manage to discount a lab strain as a source of the human infection? Were the closest Yunnan viruses not kept in the Wuhan Institute? Did anybody sample civets or other animals sold in the Wuhan market for coronavirus? (The market, of course, is closed, so this may be impossible.) Why do the authors consider the civet cat (palm civet) the most likely intermediate host of the virus? They cite this paper, showing a near-identity of the human and palm civet virus, but do they have similar data from other mammals?

I am not an expert on the various theories of transmission of cornavirus from bats (the most likely origin) to humans, but offer this for your delectation.


Lytras, S., W. Xia, J. Hughes, X. Jiang, and D. L. Robertson. 2021. The animal origin of SARS-CoV-2. Science: DOI: 10.1126/science.abh0117

34 thoughts on “The origin of the coronavirus

  1. I don’t really see any discrepancy between the genetic relationships that the paper outlines, on the one hand, and the involvement of the Wuhan virology lab in the release of the virus into the human population on the other. The paper, so far as I can tell, doesn’t directly address the question of whether there’s reason to believe that the virus that now infects human beings reflects the kind of gain-of-function research that there is reasonable evidence for at the Wuhan lab. If the two issues are orthogonal, then we are still in the dark about the exact storyline behind the origins of the pandemic, no? Or have I missed something (always, alas, a very good possibility….)?

    1. As I understand it, the Wuhan lab looked at potential human infectivity of virus genes but not of whole viruses obtained from wild animals. This was not, so far as I have been able to determine, “gain of function” manipulation that could give rise to a virus with enhanced human pathogenicity. In any case, the SARS-COVID-2 virus seems to be a generalist and not a specialist enhanced or designed to attack humans. Some of the reported hosts include ermine, weasels, otters, tigers, leopards, and lions. Late last year it shut down Denmark’s ermine fur-coat industry. I saw a news report a few days ago claiming that Covid-19 antibodies had been found in deer in the US.


        In the absence of any supporting evidence whatsoever it is irresponsible and unscientific to give any credence to the lab leak conspriracy theory, especially since the scientists at WIV have denied that they were working with SARS-CoV-2 and denied that any of their workers fell sick with COVID-19 before the pandemic started.

  2. The civet is a very interesting piece of the story. They are not only pets but a source of meat in that part of the world. Wet markets and all the rest. It is no wonder we continue to see more virus from southeast Asia.

    On a connected story, I told my barber to get vaccinated after finding out she had not done so. A few weeks later she got it and was out three weeks recovering. I just talked to her today and she admitted she screwed up.

  3. That is an odd tree. It looks like part of it is a cladogram (all the upper branch segments have the same length) and other parts of it have informative branch lengths (near the tips of the branches). Maybe an expert can weigh in on this aspect of the tree?

    1. I think its a cladogram, where the branches mean nested relationships. The branch lengths don’t allude to amounts of genetic change, if that is what you are wondering. The branch lengths are to line up all the branch tips to get the names all lined up.

      1. No Mark, it is an odd mix. As I said, the major clades along the “spine” of the tree have steps that are all the same length, but the branches near the tips vary in sizes and are much smaller. In a normal cladogram all the branches are multiples of a basic length.

  4. Regardless of the lab thesis, this report suggests reasons to brace for further outbreaks of yet unknown viruses.
    We know the previous administration was warned of the possibility but laughed it off and later shrugged off early infections.
    Let’s not repeat that folly.

  5. How many covid cases in U.S. were spread by Chinese? Free Trumpy Bear to first RepubliQan to answer. Hint: it’s less than 100.

  6. “Why do the authors consider the civet cat (palm civet) the most likely intermediate host of the virus?”

    Cats do seem to be particularly vulnerable to COVID based solely on all the reports of cats in zoos and the like testing positive. Seems they get over it quickly. Googling uncovers a bunch of people saying that pets can catch it but don’t transmit it. I don’t have any proof against this but if they get sick and breathe, it seems likely they can transmit it.

  7. COVID-19 is COronic VIrus Disease 2019. It’s the name of the illness.

    SARS-CoV-2 is the name of the virus.

  8. (no expert here but) I’m very much for the civet cat vector – recall it was the reservoir of the previous SARS 20 years ago. Further, genetic experts have (almost to a person) said the lab theory is “very unlikely” as an origin due to the fact man-made/man-screwed-with DNA leaves “large fingerprints.”

    Finally the (purposeful) lab theory is pushed by all sorts of weirdos and cranks, none of whom seem to have any qualifications in this area. I’m going with the experts.

    Further, and I wrote an article about Taiwan trying to phase out traditional Chinese woo-woo – while the PRC promote it – people in China eat bats for “medicinal” reasons, not protein or taste apparently. Fake medicine (TCM) actually got us into this mess.


  9. Will this information on the origin of SARS-Cov-2 damage the appeal of comic books and movies featuring Batman? Come to think of it, what sort of audience does DC comics have in China?

  10. I’d like to see an alignment between the top five in the cladogram (through RaTG13), and some mention of the % identities. Drilling into supplementary material all I got was a table listing the accession numbers. Are differences spread across the genomes or clustered in just a few areas. There’s already one virus known with (I believe it was gene 10) 100% identical to SARS-CoV-2, name of which I didn’t retain. Are there more in these closest three with 100% pairwise-identical genes?

    And of course the all-important similarities/differences in the Spike. Did I miss a reference to all of this somewhere?

    1. Yes, see my comment 4. Tree topology isn’t the only thing that matters. We need genetic distances to evaluate the tree. Unfortunately this is a common flaw in DNA work these days.

      1. Yep, I saw yours too. The more I think about this one, the more it seems like a sociology paper. The angle about the pork shortage is fascinating, but it’s not what I was expecting from the title. Maybe with being classified as a Perspective, different criteria apply?

        Also, it would be interesting to know if the % silent mutations are as expected from random mutagenesis, or if silent mutations are way below expectation. I suppose if none were silent we would have been reading an entirely different paper.

  11. I’m not impressed. The fact that RaTG13 “seems too distant from the human coronavirus to have been the source” is irrelevant. That’s exactly what you would expect if it was manipulated through gain of function research.

    1. I think you need to drill into ref 5 to get the background on that half-sentence summary first. IIRC, only ~30-base snippet of RaTG13 existed before COVID-19 emerged. When SARS-CoV-2 was sequenced, that snippet proved the closest match in the databanks, and it was then that a sample of RaTG13 was hauled out of storage and sequenced.

      Plus, there are now three examples of viruses with closer matches.

  12. “Did anybody sample civets or other animals sold in the Wuhan market for coronavirus? (The market, of course, is closed, so this may be impossible.)”

    This is the aspect I’m most curious about. Whole-genome sequencing has rapidly advanced in the last few years and it has been put to good use in studying COVID. As the phylogeny itself shows in the sample names, WGS of human, bat and pangolin SARS-CoV-2 variants were obtained as early as 2019 and 2020. It may be impossible to find the exact civet individuals used in the wet market but it seems like it would be really easy to find wild civets (and other putative species involved) and obtain genome sequences of their SARS-CoV-2 variants (if they have them). It is mildly frustrating that I’ve read papers released over a year ago that simply handwaved on this issue and said the virus probably transferred from a bat to some other animal and that this brand new paper seems to do the same thing. I’m aware of no positive evidence SARS-CoV-2 leaked from a lab but this inaction makes me skeptical it’s from nature. Also, the paper the authors cite about the near identity of human and civet SARS is from 2005 so it’s plainly about SARS-CoV-1.

    W. Benson in this thread brings up an interesting point though. I have heard of the SARS-CoV-2 transmitting from humans to other animals. I haven’t read into that much but if the “human” variant is a generalist that can effectively infect many animals then that would provide some explanation of why it’s been so hard to find the vector that originally transmitted to humans.

  13. One can only hope that solid research, such as this, will put the kibosh on Loury and Weinstein’s seemingly egoistic campaign of mis-/dis-information.

Leave a Reply