Natural COVID crowd miss again
"On reading," by Simon Wain-Hobson, is a weekly discussion of scientific papers and news articles around gain of function research in virology.
Since January 2024, Dr. Wain-Hobson has written weekly essays for Biosafety Now discussing risky research in virology. You can read his entire series here.
On reading Dynamics of natural selection preceding human viral epidemics and pandemics by Jennifer Havens, Sergei Kosakovsky Pond, Jordan Zehr, Jonathan Pekar, Edyth Parker, Michael Worobey, Kristian Anderson and Joel Wertheim, Cell 2026; S0092-8674(26)00171-6.
Scientists get a glimpse of how new pandemics are made by Carl Zimmer, New York Times, March 29, 2026.
A ‘nail in the coffin’ for the lab-leak theory? by Mark Honigsbaum, The Telegraph March 19, 2026.
This is another paper from a group of scientists who have been striving to show that the COVID-19 virus is of natural origin. This is an unusual scientific endeavor given that people are not cooperating – WHO’s opinion (WHO wordplay on COVID origins). There are no sister viruses around just distant cousins so it’s possible to mess up.
In the paper’s abstract there is an odd line: We conclude that extensive pre-zoonotic adaptation is not necessary for human-to-human transmission of zoonotic viruses.
The thing about evolution is that when confronted by a change selection occurs on variants within the population of animals, plants or microbes allowing adaptation to the change. If unsuccessful, competitors ever present get the upper hand.
The microbe will not pre-adapt unless confronted by that change. In other words, a virus cannot adapt to human-to-human transmission without first getting into humans. More specifically, a coronavirus in a bat or a mammal cannot pre-adapt to humans. Humans will always be different. You can write the sentence, but so what?
In the introduction they write that some viruses circulating in the natural host reservoir already have the ability to infect and transmit between humans and do not require additional adaptation to cause zoonotic epidemics.
On reading was but a few years into his career in virology when he learnt that the best model for man is man. Even HBV in chimpanzees doesn’t fully recapitulate all aspects of the human disease. There are large numbers of differences between humans and chimps, something that has been known for 25 years or more. For example, human chromosome 2 was formed by fusion of two ancestral chromosomes in the chimp.
Note that they are concluding a negative, something is not necessary, which is always fraught with problems. Why didn’t the dog bark? The most plausible explanation is that it knew the intruder, dixit Sherlock Holmes. But the intruder could have drugged the dog, it had laryngitis, was deaf or distracted by canine social media. The list is long. Scientists put a premium on proof, especially for publications in a top journal such as Cell.
The paper presents no new data, although it provides a reworking of a program published a decade ago. It’s voluminous with the multiple threads coming together at the end. So, what’s the idea?
They feel it should be possible to differentiate between the evolution of a virus in its natural environment to that in lab culture. This is to test the hypothesis that if the SARS-CoV-2 virus leaked from a Wuhan Institute of Virology lab it would carry vestiges of lab culture in its genome. These would show up in the ratios of certain types of mutations, at least that is the postulate.
They compare examples of natural selection of several viruses in the wild and viruses adapted in tissue culture. Fine, but going down this route, they can at best establish correlations, no more.
If the virus spends a short period adapting in a novel host, this evolution is unlikely to result in a detectable change in selection. Intuitively, for the method to be powered, it is necessary to have a selection change that is sufficiently large and an evolutionary time that is sufficiently long to accumulate detectable substitution patterns. A single or a few substitutions, even if they are adaptive, will likely be missed.
Early on we learn that their method is only sensitive to genetic changes that are sufficiently large and an evolutionary time that is sufficiently long. We have no idea what the precursor to the SARS-CoV-2 is although, once upon a time, it was a bat coronavirus. As we are never told how large or long are, we have no idea if these conditions are met.
We’re given sections devoted to the 2009 pandemic flu virus, Ebola and Marburg filoviruses, monkey poxvirus, coronaviruses of the SARS family and the 1977 Russian flu pandemic virus that inadvertently emerged from a lab by mechanisms not fully known. For all of them purifying selection - rooting out of the less fit genomes - dominates. The extent of the purifying selection varies but this not new.
Their analyses are developed to measure changes in selection pressures or to use their words intensification or relaxation of selection compared to some reference. Changes in selection pressure could signal host changes of lab culture.
If you look carefully there are anything but comparable numbers of mutations for the background, stem and outbreak groups. Only for the former are there many mutations which explains the relative stability in purifying selection. For the others there is greater variation in part due to fewer mutations sampled. They bundle everything into a single parameter called selection intensity as if evolution of viral genomes can be represented by a single number.
Fitness and therefore evolution are multigenic.
Two further technical points. First, it is fair to conclude that host switching occurred along the branches called stems in the figures. Accordingly, they harbor genetic changes in two species and so are composites. As nobody knows a posteriori where along the stem zoonosis occurred, there is no way of allowing for this. This renders the findings unreliable and undermine statistical inferences made from calculations.
Second, they use multiply aligned genome sequences so important features like the unique furin cleavage site in the SARS-CoV-2 Spike gene/protein are eliminated.
As the raison d’être of this work is exclusively COVID origins, what do they say about SARS-CoV-2? In the results our analysis of selection on the stem preceding SARS-CoV-2 emergence… found no evidence of intensification or relaxation of selection compared with selection of the bat host reservoir. Hence, we find no evidence to suggest SARS-CoV-2 experienced prolonged selective pressure in an environment different from related bat viruses prior to its emergence in humans.
They find no evidence.
• Maybe the averaging the effects of genetic drift in one species followed by rapid selection in a second species along the stem scrubbed the signal.
• Hypothesis: the furin cleavage site is the biggest determinator allowing increased replication and transmission of SARS-CoV-2 between humans. As it is deleted by making multiple alignments, the negative findings across the rest of the genome are unsurprising. Discuss.
Remember also that using serological techniques US and Chinese researchers showed that bat coronaviruses spillover into humans. They estimate that a median of 66,280 people are infected with SARSr-CoVs annually in Southeast Asia. To make antibodies means the virus replicated well in humans, just not enough to ensure onward transmission.
These spillover viruses have solved a great deal, but not the human-to-human transmission issue. And that is huge. Therefore, adaptation is necessary. Indeed, this is exactly the issue Drs Fouchier and Kawaoka addressed for avian H5N1 flu spillover infections in humans. Their aerosol ferret transmission work on H5N1 ignited the dangerous gain of function research controversy back in 2012.
The authors analyze six mammalian viruses that have been passaged in tissue culture, some extensively, others less so. The results were mixed. For three cases there was evidence for less purifying selection – relaxation of selection - which is as understandable given that there is no place more hostile for a virus than inside an animal with acute, innate and adaptive immune systems. For the mumps virus there was very little change.
For two cases, mouse hepatitis coronavirus and human H2N2 influenza virus, the results were consistent with adaptation to tissue culture during passaging.
In short, for the six viruses propagated in the lab there was considerable variation in the outcomes.
What have they to say about SARS-CoV-2? These sensitivity analyses indicate that changes in selection intensity due to human-to-human transmission within a single viral genome lineage are relatively weak compared with those induced by laboratory passage, and a combination of multiple sampled genomes over many months is needed to reliably detect this change. The six examples were anything but uniform as the words in bold (added) might suggest. If your references are mouse hepatitis virus and H2N2, OK. But mumps virus? There was little change.
The authors have no way of knowing which is the more appropriate reference virus. Note they only make a comparison, a suggestion that… no proof. And that’s it.
To quote Cyrano de Bergerac, Ah! non! c’est un peu court, jeune homme! (Ah! no! that’s a trifle short, young man!). The manuscript clearly encountered obstacles during its 13 month review which was unusually long. Yet Cell went with it.
The fact that the virus evolved quickly from the prototypic strain to the alpha, beta, delta strains and onwards to the much more transmissible omicron strain – something the authors pick up - at a time when there wasn’t worldwide herd immunity means the Wuhan strain was not pre-adapted to Sapiens. So what does pre-adaptation mean?
Pre-adaption is in the mind of the authors. It’s surprising and sad that the referees and Cell editors went along with this thin gruel. The hound of the Baskervilles surely should have howled.
This article made the news, notably in the NY Times authored by science writer Carl Zimmer. The researchers found no peculiar changes in SARS-CoV-2 before it jumped into humans. It gained mutations as it spread from bat to bat, just like other bat coronaviruses did; only after the virus emerged in humans did it undergo a marked shift. Within a year, radically new variants were evolving, with mutations that made them exquisitely well-adapted to humans.
Apparently, the precursor virus was banging around from bat to bat and then went haywire, radically new variants were evolving, with mutations that made them exquisitely well-adapted to humans, as the COVID pandemic exploded. Odd, no, for a virus that was pre-adapted?
The article finishes with the take home message that Zimmer has been pushing for a while: “Most of the peer-reviewed scientific evidence supports this hypothesis,” the scientists noted. Dr. Wertheim has helped collect some of that peer-reviewed scientific evidence. The new study adds more evidence in favor of an animal origin, he said.
The quip Most of the peer-reviewed scientific evidence was handled in the recent essay (WHO wordplay on COVID origins). This paper doesn’t provide any new evidence in favor of anything. It provides a handwaving suggestion that doesn’t stand up to scrutiny.
To be fair Carl Zimmer, like so many, doesn’t have the wherewithal to handle the fine details being bandied around. That’s why peers are asked to review specialist papers. As we have seen, peer review can break down (High risk research and public trust).
The Telegraph story takes a resolutely anti-lab leak stance despite noting that Six years on from the biggest pandemic of modern times, we are no closer to knowing how and where the virus first spilled over into humans. There are some cheap shots at the NIH Director and Matt Ridley a science journalist, not a virologist, and a leading proponent of the lab-leak theory.
We read that The row comes as a new study in Cell calls into question one of the central planks of that theory: namely, the claim that when SARS-CoV-2, the virus of Covid, emerged in Wuhan it was suspiciously well-adapted to people.
In other words, the virus was just waiting for the right opportunity to infect people which is utter nonsense. A sentence without meaning. The lead author of the paper is down as saying “That absence of evidence is exactly what we would expect from a natural zoonotic event – and it represents another nail in the coffin for theories invoking laboratory manipulation.” Ah, the irony of absence of evidence. Remember, the authors themselves deleted the furin cleavage site from their analysis.
The piece ends with Unless and until evidence emerges to suggest otherwise, that remains the most parsimonious explanation. Hmm. Before 1977 all pandemics were natural. However, if you take the last 50 years as reference period – around three-quarters of the world’s population was born since - 1977 pandemic flu resulted from human intervention, 2009 flu was natural while 2020 COVID is hotly contested. The parsimonious explanation no longer works.
Note that Honigsbaum, a medical historian and science journalist with a PhD in history so not a virologist, asks readers to think differently to Matt Ridley, a science journalist, not a virologist with a PhD in zoology who at least has written a book explaining his reasoning.
Conclusions
• A multitude of thoughts and data sets relying on many assumptions bucking much settled knowledge in evolution. In a top journal this paper can harm virology, and consequently public trust.
• Six years on the natural origin crowd miss the target once again.
• Science news journalists, stick to facts. If you can’t master them, don’t pen anything. Your editor won’t be happy but there is far too much nonsense already out there – in peer reviewed papers and elsewhere. Above all, remember, peer review starts with publication (WHO wordplay on COVID origins). A hot of the press paper is not settled knowledge.
• Artemis II didn’t go to the moon and back based on the equivalent of Cell papers like this.
Aside 1
Talking of chimps, overall, their genome is about 99% identical with that of Sapiens. Suppose our nearest neighbor genetically speaking was the gibbon whose genome is about 96% identical. We’d be missing the orang-outans, gorillas, chimps and delightful bonobos.
Accordingly, it is fair to say that as the SARS-CoV-2 and RaTG13 genomes are 96.4% identical we’re missing a great deal.
Yet six years later, greater than 25 million dead, nobody has bothered to fill in the gaps, aka the thousands of SARS2 like bat coronaviruses that are out there in the wild. The ones causing 66,280 spillover infections annually in SE Asia. It’s not normal. At all.
Aside 2
In the results section they write that Genetic and epidemiological evidence support the hypothesis that the virus preceding SARS-CoV-2 briefly circulated in an intermediate host sold at the market. What genetic data? What epidemiological evidence? What intermediate host and for how long? The sources subtending these lines provide suggestions but not answers. This is a paper in Cell so things need to be rock solid from A to beyond Z. Correlations are not good enough. As to the Huanan market, maybe it functioned like a superspreader event but there is no evidence that it was crucial for the virus to jump from of an as yet unidentified intermediate host to humans.
Aside 3
This paper was covered in an April 28, 2026 podcast by Dr. Vincent Racaniello, a professor of virology at Columbia University, NY. As the work was described as neat around 56 minutes, apparently little thought was given to the paper. Dr. Racaniello has been a staunch disparager of the lab leak hypothesis from the get-go, just as he’s been a supporter of DGOF for even more years. It appears based on the position that virologists know best which, sadly, seems impervious to data. So, for the record.
Aside 4
The recent hantavirus outbreak on a cruise ship is caused by a strain of Andes virus, something we have discussed before (Andes virus).
Always on the nail! Thank you.
The directions of the BioStrategic Domain and the AI&Robotics domain are very disheartening. Capital is being lured by the dystopian fair-weather thinking of authoritarian states such as PRC...even as they send chilling warnings of all out war with the free world without end.There is some Hope.Excellent work here from Stanford etc.
https://www.hoover.org/research/seeing-through-molecular-curtain/The
The Biosensor direction, especially when deployed in a BioStrategically significant setting such as a conference that is shaping best practice biosecurity, creates a powerful opportunity for collective accountability. The stated target diagnostic band, with only three viruses, is currently narrow, but broadening this window towards collection of airborne traces of bio fragments linked to novel and dangerous Dual Use <<Trust but Verify>> research of concern can help forensically triangulate bioinformatic traces found retrospectively in raw reads from available databases.When deployed in a preemptive and realtime way this is more in line with the expectations from other ontological extinction level weapons grade research such as in the nuclear weapons domain.
https://www.encounterbooks.com/books/the-code-as-witness/?srsltid=AfmBOoonD9n5peEgkQzLAUI8hbr8dRcISqkPyKf4IYJmLtwST37uTRZyhttps://www.linkedin.com/feed/update/urn:li:activity:7458577596436619264/?dashCommentUrn=urn%3Ali%3Afsd_comment%3A%287461174077022474241%2Curn%3Ali%3Aactivity%3A7458577596436619264%29