Too Much
"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.
Welcome to 2025. These essays continue to discuss risky virology being performed in academic labs that generates novel human viruses. Using quotes — given in italics — from published papers and online sources the essays show how some studies are unable to generate meaningful results. Or they highlight inconsistencies and incorrect statements in their own words. Occasionally it’s breathtaking.
References are embedded in the text, previous essays are referred to in brackets while a few tangential remarks that might slow the flow, are pushed aside.
Let’s start 2025 with a DURC or research of concern paper on bird flu viruses. The idea is simple. All you need to know up front are two things. One, 1918 Spanish was the mother of all flu pandemics – around 50 million dead in a world population of around 1.9 billion, aka over four times smaller than today. Two, human and bird flu A virus genomes are made up of 8 different segments.
The idea was to make a bird flu virus using genome segments from circulating viruses as close to the Spanish flu virus as possible. Why?
Wild birds harbor a large gene pool of influenza A viruses that have the potential to cause influenza pandemics. Given this the possibility exists for a 1918 virus-like avian virus to emerge in the human population as a pandemic virus; however, the likelihood of such an event remains unknown. To assess the risk of emergence of pandemic influenza viruses reminiscent of the 1918 influenza virus, they trolled virus databases with computer programs that have been around for years, no AI needed, and found the nearest bird lookalikes to Spanish flu for all 8 segments.
The 8 segments differ from the bona fide 1918 virus by just over 110 differences at the protein level which is huge for a specialist. They were able to make virus from these 8 disparate gene segments generating what they call a 1918-like virus. While this may sound surprising to some, it is the realms of the possible. Not predictable, but not stunning.
The ferret is considered the best current model for influenza virus infection because infected animals exhibit symptoms that resemble those of humans infected with influenza A virus. When injected into ferrets the 1918-like virus caused weight loss unlike an authentic bird virus although not so much as the original 1918 virus. They note that, For an influenza virus to cause a pandemic, it must achieve efficient human-to-human transmission. Thanks, we had worked that out some time ago.
They did the experiment to ascertain ferret to ferret transmission and it failed. By contrast, the 1918 virus was transmissible meaning the experimental set up worked properly.
Written in 2014, the first words of the title of this paper are circulating avian influenza viruses. We all know that flu viruses come and go with most becoming extinct - think of the waxing and waning of seasonal flu strains. Or across evolution where most species go to extinction. We are not told if any of the 8 viruses from which a genome segment was derived are circulating today. The question is valid for the segments were found among duck viruses isolated between 1975 to 2002, aka some 12-37 years before the current work was performed.
An engineer uses everything at their disposal to solve the problem. Present and past information, everything is fair game. Evolution doesn’t work like that. It works with what is at hand. Extinct genomes are not at hand.
If they had used duck viruses circulating around 2012-13 when they were doing the work, then at least they would have been addressing the question. Instead, we get some back-from-the-past creation made possible by computer screening.
This is not a 1918-like virus. The logic is wrong. The approach was irreal from the outset whether it delivered or not. And nobody saw this during grant review?
But they’re not finished.
They go on to exchange a couple of 1918-like segments for those from the bona fide 1918 virus. Of the resulting viruses two were transmissible to 1/3 and 2/3 ferrets by the aerosol route indicating that the 1918 virus has something the pseudo-1918-like virus doesn’t. This is hardly surprising given that it was, oops is (now that Spanish flu has been resurrected by CDC and NIH labs (1918 and all that), a human virus whereas the pseudo-1918-like virus is uniquely a mosaic of duck genome segments.
They went out of their way to find mutations that increased the pathogenicity of the 1918-like mosaic viruses and indeed found some. Surprisingly, while pathogenic in ferrets directly injected by some of these viruses, none of them killed animals following airborne transmission.
We’re left with a collection of experimentally defined mutations in a totally novel virus, conceived with flawed logic, of unknown potential. This happens, there’s nothing wrong with that. But then, why flog a virus that doesn’t transmit by aerosol?
The answer can be given by many a senior researcher. It goes something like, ‘Well, we’ve invested so much time and effort in this project, we must drum up some positive results, bang everything together and get a paper out. After all, X is a student while Y a post-doc. Both need papers to advance their careers. Furthermore, we need a paper to show the funding agency.’ This explains the experiments from page 693 of the paper onwards.
We have to accept statements like we demonstrated that acquisition of only a few amino acid substitutions can confer respiratory droplet transmission to 1918-like avian influenza viruses in a ferret model, suggesting that the potential exists for a 1918-like pandemic virus to emerge at any time from the avian virus gene pool.
The first quote from the paper used above was To assess the risk of emergence of pandemic influenza viruses reminiscent of the 1918 influenza virus. Yet the only precision we have on assessing this risk is the potential exists… to emerge at any time. We knew that a while back. And you needed support by National Institute of Allergy and Infectious Diseases Public Health Service research grants, by RO1 AI080598 and R56 AI099275, by ERATO (Japan Science and Technology Agency), and by the Strategic Basic Research Programs of Japan Science and Technology Agency, Japan to tell us that?
Practically speaking, the data about this 1918-like virus is available to everybody. They are of unknown potential for humans should there ever be a lab leak. Furthermore, all of this work is unfalsifiable so we can’t use it. Can we insist that the virus stocks are destroyed to reduce the risk of a storage accident, based on the logic that destroying smallpox, rinderpest and poliovirus stocks makes the world a safer place? (Rinderpest).
A New Year’s gesture and all that. Thanks.
Aside 1
Talking about viruses going extinct you may be amused to know that the vast majority of primate species, an estimated 96%, went extinct? Furthermore, the average time for each species was around 1 million years. Only a million years. Yet Homo sapiens diverged from Neandertals around 500,000 years ago. That leaves us about the same amount of time. All that for that?
Aside 2
Notice that the way to save the project was to generate a ferret transmissible virus out of one that doesn’t transmit. That is a Gain of Function of great concern for it doesn’t help virology even if it generates a paper for the lab. An unquantifiable risk for society versus an identifiable benefit - a scientific paper. IMHO of course.
Yes, destroy the stock of these extinct viruses. Setting aside gain-of-function, lab safety and monitoring, this alone would reduce the risk of a bio disaster. Any amount matters astronomically.