An evolutionary void

"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.

Malachite, not map contours

On reading Is pathogen prediction possible? by Arturo Casadevall, PLoS Biology 2025; 23: e3003162

---

This ‘Perspective’ starts with Casadevall being taught in medical school that retroviruses were not pathogenic for humans and that coronaviruses were at most a winter nuisance causing colds and rhinorrhea. HIV and the SARS viruses overturned that. Clearly, history and experience with infectious diseases do not predict future threats, thus raising the fundamental question of whether prediction is possible for future pathogenic microbes.

Coming from a staunch supporter of dangerous GOF virology which was touted by Drs. Fouchier and Kawaoka as helping predict the next pandemic such a remark can’t be passed over. Both sentences are reminiscent of that phrase from investment bankers, Past performance is no guarantee of future results.

The rest of the article can’t be passed over either. We learn that For viruses, there are 10–100 pathogenic species per mammal from an estimated 10⁷ to 10⁹ viral species. While such numbers are highly uncertain and almost certainly underestimates, the point is that the proportion of microbes with the capacity to cause disease in any human is very small. This in turn implies that the characteristics needed for virulence are rare and/or that humans have formidable immune defenses that can prevent the overwhelming majority of Terran microbes from establishing themselves in the host.

Nope. It’s not the absolute number of viral species on the planet that counts. And while 10-100 pathogenic species for man is more than enough most of the 10⁷ to 10⁹ viral species actually infect bacteria. So 10-100/10⁷-10⁹ is a meaningless ratio. Accordingly, the conclusions are equally meaningless. The relevant question to Casadevall’s wonderings is what fraction of human microbes cause disease?

Notice than the last sentence given above offers two competing hypotheses, or even a mélange of both. Odd that the next sentence is so trenchant: Given that the capacity for virulence among microbial species is rare… Someone should have proofread the manuscript.

What was AIDS in the early days? Lists of microbes isolated from patients, many of which were not pathogenic in people without HIV. So yes, the immune system is a phenomenal force for knocking down microbes. Studies on a strain of simian immunodeficiency virus in rhesus macaques showed that genetic lesions in genes showed that all contributed to viral growth, one a thousand fold, another merely a factor of just two. Virologists would say the virus is on a fitness peak where the only way is down. They are so fit that 10¹⁰ – 10¹³ virions of HIV, HBV or HCV virions are cleared from the human body per day.

Why so many? To be transmitted. Without transmission it’s a spillover infection. An evolutionary dead end.

The thing is the more a virus grows, the greater the chance of transmission. Indeed, transmission can sometimes be mapped back to a single virion. This makes sense given the virological mayhem described in the numbers above.

By and large, the more a virus grows, the more the disease, virulence in Casadevall’s lexicon. However, let’s just take two examples to realize that transmission and virulence can be uncoupled. HIV is mostly transmitted during asymptomatic primary infection although the horrible disease called AIDS appears 5-8 years downstream. Or COVID. Remember in the spring of 2020 the shock of learning that the virus was transmitted during the asymptomatic period preceding acute infection. Specialists said, ‘that changed everything’.

Which is the more important of the two – transmission or virulence/disease? The ambiguous word is important. From the human perspective disease is the pain. The entire biomedical enterprise is designed to cope with disease. So naturally, disease looms large our lexicon. For the microbe transmission is paramount. That the host dies is not an issue just so long as it is transmitted before the death of the host.

Casadevall doesn’t mention the word transmission once in his perspective and so deprives the reader of crucial knowledge. This is not another endless academic sparing between two elderly gents. It’s a seriously big omission that shouldn’t happen.

Let’s do a thought experiment. Consider a highly pathogenic virus that kills 60% of those infected but which is not very transmissible giving rise to an epidemic of 400,000 infections. ALSO might note a whiff of the Western African Ebola epidemic. Likewise consider a highly transmissible virus that infects at least 1 billion people but kills only 1%. Shades of COVID? It’s clear that the latter virus is way more dangerous as it would kill 100 million, whereas the highly pathogenic virus would kill 240,000. Obviously, disease is harrowing, but you must never forget transmission.

A transmissible agent doesn’t have to cause disease. Pathological consequences like coughing, sneezing, vomiting, diarrhea or skin pustules all help transmission for the microbe is omnipresent in these ejections. Droplet sizes range from < 1 μm to 1000 μm, and in typical breath there are around 100 droplets per litre of breath. So for a breathing rate of 10 litres per minute this means roughly 1000 droplets per minute, the vast majority of which are a few micrometres across or smaller. Coughing and sneezing just increase the number of droplets and with them the respiratory virus.

Yet some viruses like the human circoviruses infect over 50% of humanity yet don’t cause disease. Blood banks don’t screen for them. Our guts are full of bacteria who live in symbiosis with us.

All the above are at odds with the remark, the capacity for virulence is such a demanding phenotype that it is rare and can be fleeting. Disease varies. And if a lowly E. coli picks up a virulence factor on a plasmid all hell could break loose.

Casadevall goes out on another limb, this time with one can anticipate that a virus from bats, a group of species that shares the mammalian class and body temperature with humans (while not hibernating), is more likely to pose a threat than a salamander virus. That is plausible but hints at mutual exclusion. Consider for a moment insect borne viruses, those that multiply in both cold blooded insects and warmer bodied humans. Their names are household words – yellow fever, dengue, Zika to name but three. Furthermore, there is the redoubtable vesicular stomatitis virus of cattle or equine encephalitis virus. It wasn’t a robust remark.

With the paragraph starting Apart from the epistemic limits imposed by emergence we learn of some of the factors creating uncertainty. For example, Unfortunately, relatively little work has been done to understand the dynamics of host-microbe relationships. Which is true. It would be good if far more lab virologists switched to studying virus replication in vivo, quantifying the tempo and magnitude of the titanic struggle between a virus and the immune system.

Titanic? If 10¹⁰ – 10¹³ virions of HIV, HBV or HCV are produced per day then ~10¹⁰ – 10¹³ virions of HIV, HBV or HCV are destroyed per day otherwise the immune system or the liver would be demolished in an instant.

Relatively few do in vivo work because it is multidisciplinary requiring knowledge of physiology, virology, immunology and mathematical modeling. Its labor intensive, more costly and takes far more time to complete compared to quick tissue culture experiments. Change here can only come about via the funders.

The perspective collapses with a forage into weather forecasting. As the latter has become better and better over the last 50 years so might it be so for pandemic prediction, for this what the subject is really about. Perhaps the field of microbial pathogenesis can take a page from the science of weather prediction and combine increased sampling of the natural world with clinical observations, experimental results of pathogenesis experiments, and historical data to identify potential threats and assign a risk probability. Perhaps. However, over the same period the weathermen have made progress, we’ve seen HIV/AIDS, the 2009 flu pandemic, COVID and a multitude of severe epidemics aka Zika, SARS, MERS, Ebola. Not one was predicted.

It is a distraction thrown out to deflect from the difficult questions.

And nowhere is this more relevant than in evolution which is replete with historical ifs and buts. No engineer would ‘design’ a human genome with up to 10% of bits and pieces of retroviruses spread across all chromosomes, not to forget bits of RNA virus genomes and mitochondrial DNA that also got lodged in our genomes. This is biology and evolution works on it. Even if certainty in prediction is not possible in microbial pathogenesis, it may be possible to construct a probabilistic framework to identify those microbes that pose the greatest threats. Such lists exist without probabilities, unless maybe, someday, perhaps are good enough for you. Such words are a fat lot of use to public health planners.

In fact, developing predictive models for the identification of future threats and testing predictions in animal models could supercharge progress in the field by allowing the identification of relevant variables and their relative importance. When a new infection emerges three things are done. Isolate the agent. Sequence the genome. Stick it the agent into as many different animals as possible to generate a ‘model’ to get some handle on pathology and test lead therapeutic molecules. This trio hasn’t changed in the last 45 years.

The same is true for the emergence a novel strain of an existing virus. Just how robust animal models of human disease is another issue and here the quip is the best model for man is man. And that’s ethically impossible. Researchers try to get as close as possible and occasionally fall into the trap of believing their models do recapitulate human disease. See next week’s essay.

Dr Casadevall wants to be upbeat, optimistic. Who doesn’t? However, he cannot contemplate that pathogen prediction is not possible which is one legitimate answer to his question. Or limited for many of the reasons mentioned above.

Post COVID pathogen prediction is part and parcel of pandemic prediction. Indeed, the term pandemic prediction has exploded since 2019. A PubMed search using ‘pandemic prediction’ revealed a whopping 19,393 papers from 2020 to present. A mere 1514 came from the period 1973-2019.

Before COVID the buzz phrase was emerging infectious disease following the CDC’s 1993 publication of the Sin Nombre hantavirus from the Four Corners region of the US. The CDC established a journal called Emerging Infectious Diseases on the back of that important discovery. From that time on, there were almost as many interpretations of the term as papers – 36,873 in fact. Suddenly genetically stable viruses were emerging while the term re-emerging infectious disease – over 13K papers - was quickly coined to cope with outbreaks of well-known pathogens.

When you learn of 19,393 papers, meaning 150K+ authors, that implies general agreement. However, it’s a dream, a goal which is perfectly valid. Yet is it a plausible or robust question when our track record is zero? OK, some success might be possible for some bacteria that evolve way more slowly than viruses.

Pandemic prediction has become a cottage industry with usage and grant money driving it. A pity to say the least that the premise of pandemic prediction, which we encountered with the Fouchier and Kawaoka papers, still goes unquestioned.

It is sad that the PLoS Biology editors didn’t spot the evolutionary void. Like transmission, the word evolution is nowhere to be found in the paper. Every day they are sizing up manuscripts and should know Dobzhansky’s 1973 remark ‘Nothing in biology makes sense except in the light of evolution.’ It should be plastered everywhere in their minds, much like the line from The Beatles song: Penny Lane is in my ears and in my eyes. Dobzhansky’s remark even has its own Wikipedia page. Penny Lane too.

Conclusions

The short answer is not today. Will it ever? The answer is in the future like all the problems scientists work on. And apart from the uncertainties Dr. Casadevall highlights, there may even be unknown unknowns. As Einstein said, if we knew what we were doing, it wouldn’t be called research.

It’s the lack of an evolutionary based analysis that leads the author astray. Evolution plays the long game where survival dominates. For viruses that means transmission is #1 not disease. It could not be otherwise.

---

Aside 1

Re HIV, transmission is mainly sustained by during the asymptomatic phase. AIDS tends to 100% if left untreated and with it death. There is no greater disconnect.

---

Aside 2

Re retroviral DNA fragments in the human genome, some have been hijacked and put to good use. Human syncytins are key proteins involved in placentation. Their origins are found in retroviral envelope proteins, those used to dock and fuse with the target cell.

---

Simon Wain-Hobson is an emeritus professor at the Institute Pasteur, Paris, from which he retired in 2021. He and his colleagues were the first to sequence the genome of HIV, and Wain-Hobson has published more than 230 papers on virology and cancer.