NOTE: On Sunday we’ll have another post about the biology of the coronavirus variants, but that one will be by Greg Mayer, and concentrates specifically on what these new variants are and what they mean for us.
*******
I’ve capitalized evolution in the title because Covid-19 offers an object lesson in evolution by natural selection, yet people don’t seem to grasp that aspect of the pandemic.
Last night, while watching the news, I was startled when a report on the new “highly spreadable” variants of Covid-19 mentioned “the evolution of the virus.” Although viruses clearly evolve, and most of us know about this, the e-word is rarely mentioned on television, perhaps for fear of offending the many Americans who are creationists. Yet those same creationists get their flu shots every year because flu viruses evolve, probably to become more transmissible as well as to avoid being destroyed by our naturally acquired immunity.
Just a few words about evolution in Covid-19 and viruses in general. First, mutation is not evolution; it is the raw material for evolution. But often on the news they refer to the appearance of a new mutant variant as “the virus evolving”. But evolution is the change in the frequency of variants, not the appearance of new ones. (That change can reflect either natural selection or random genetic drift.) You have to have mutation to have evolution, but they are not the same thing.
Most of the data we have about genuine evolution in the Covid-19 virus is the rise in frequency of new variants, probably by positive natural selection. New variants like the “British strain” or the “South African strain” seem to be increasing in frequency, and it’s apparently not due to random genetic drift because the increase is observed in several places. These parallel changes in frequency represent a good sign of positive selection. This selection, as the media should emphasize, occurs because virus variants that are more readily transmitted from person to person are the ones that come to predominate in the population of viruses—the proportion of variants among all human hosts. If you aren’t easily passed on to other people because you don’t get into the respiratory system as easily—or don’t “take” when you are transmitted to another person—your frequency will drop.
Thus we will expect to see, from time to time, the increase in frequency of new viruses that take over the population. These won’t necessarily be viruses that are more dangerous, but simply those that are passed on more easily. There’s no clear reason why transmissibility should be correlated with “virulence”, defined by me (and most people) as how sick an individual gets when infected. There are some cases where the two factors might be positively correlated, but that’s not always the case, and sometimes they’re negatively correlated (see below). Thus people shouldn’t confuse how easily a virus is passed on with how sick it makes you. Of course, the more easily a virus is passed on, the more people in the population will get sick, but that’s not the way most people measure “virulence.”
I should add here that we know of cases in which viruses that are easily transmitted, and are selected for that property, can evolve reduced virulence. Myxomatosis (a viral disease of rabbits) was introduced into Australian rabbits to kill them off, for the invading species was destroying vegetation and displacing native wildlife. But the virus evolved to become less dangerous, because strains that killed off rabbits very quickly were strains that weren’t easily passed on from rabbit to rabbit. (This is “positive natural selection for reduced virulence.”) Thus the Aussies have had to introduce new viruses to keep the rabbit population at bay.
Viruses can also be selected to avoid the kind of “natural” immunity acquired by our past encounters with viruses, including non-Covid-19 coronaviruses. Each person will have, even before vaccination, a natural level of resistance to coronaviruses based on having encountered them before and having various degrees of immune reactions. Any variant that is able to evade this “natural immunity”—by withstanding the mix of antibodies in the population—will persist, and that too is a form of selection. Thus viruses can experience positive selection to avoid natural immunity, and these, too, will increase in frequency via natural selection. But those viruses don’t necessarily make you sicker—although they can.
The reason why the flu virus H1N1 (not a coronavirus) spread so rapidly and also made people so sick in 1918, killing between 20-50 million, is that that virus almost certainly jumped into humans from another species i, and spread rapidly because we had no natural immunity to a virus like that (there would also be the usual selection to increase spreadability). The reason it was so virulent, however, was probably not connected to its spreadability. Instead, it made us sick because that virus evolved in another species where it had probably had been “tamed” by selection (if you kill your host immediately, as in myxomytosis, you’ll be selected against); but there had been no “taming” in humans. Ergo the virus went wild when encountering non-adapted human genomes. The result: a virus not evolved in humans wreaked havoc when entering the human body.
The reason we get a new flu shot every year is probably not because new strains of flu viruses are resistant the vaccinations we get (which are designed to produce antibodies against four strains), but because they’ve managed to overcome naturally occurring immunity. (Not everyone gets a flu shot.) It’s thought that the major reason for the rise of new flu strains every year has little to do with overcoming the specific vaccines created to counteract flu.
Finally, viruses can also increase in frequency if they manage to avoid the antibodies generated by manufactured vaccines; these viruses are novel enough that they aren’t affected by our acquired antibodies. This is what people are really worried about with Covid-19. Will my Pfizer shots protect me against these new variants? And yet the selection for the “South African” and “British” sprains, based on their ability to spread faster, doesn’t seem to have been accompanied by a much greater resistance to the vaccines that are being used. Again, spreadability wasn’t accompanied by increased virulence.
I hope I haven’t confused people too much; the area is quite complex. My main points are these:
- What we see with Covid-19, and many other viral diseases, is evolution by natural selection. Evolution deniers have no explanation for these new strains of Covid-19 that are spreading so rapidly. And the media should spend more time emphasizing the evolutionary aspects of the pandemic.
- The kind of selection operating on viruses is complicated. There is selection to be spread more easily, selection to evade naturally-acquired immunity, and probably some selection to evade immunity acquired by vaccinations. (The latter resembles the ability of new bacteria to evade widely used antibiotics.)
- Because there is no clear connection between a virus’s rate of spread and how sick it makes you, we shouldn’t worry so much about individuals getting sicker as new strains spread; but we should worry about more people dying because more people in a population get sick when a virus spreads more rapidly.
- We shouldn’t get overly alarmed about the new, more spreadable viruses being able to evade the defenses we acquire from vaccinations. Both in theory and now in practice, the vaccines we use appear to work quite well against more spreadable Covid-19 variants. But that doesn’t mean that new vaccines might not be useful. In fact, I think that eventually, along with our yearly flu jab (flu isn’t a coronavirus), we’ll have to get a coronavirus vaccine as well to “top us up.”
This is a complicated area, and I’m not trying to offer medical advice beyond “get vaccinated, practice social distancing, wear your masks, and wash your hands”. You should do that even if you’ve been vaccinated, too, as we don’t yet know whether vaccinated people, themselves resistant to illness, could still get a form of asymptomatic infection that could be passed on to others through, say, expelling viruses lodged in your respiratory tract. All I’m trying to say, but then got bogged down in my pedantry, is that the Covid-19 is a good example of evolution by natural selection.
h/t: Jim Bull
SUB
Mutation proposes, evolution disposes.
I dislike the public communication about the new strains. Here in Ontario, government communications are pretty silent. On the news, they are speculative. We don’t know why these new strains seem to be more transmissible (if that’s the case because I don’t know if we know that either)….no one seems to clearly communicate this to the public and health precautions where I am have not been updated. Clear messaging is so important because most people don’t understand evolution and when they get confusing communications they tend to distrust all public information & we already know where that leads.
Very nice summary; I’ve often thought that such viral diseases, and particularly this pandemic, make an excellent object lesson or example in the nature of evolution through natural selection. I particularly think point 3 is important for the public to understand, both to avoid alarmism AND at the same time to reinforce the fact that precautions are, if anything, more important than ever.
I often worry about the metaphorical description of evolution as having a teleological nature (as in saying something evolves in order to, for instance, evade immunity) but I think you did a very good job making it clear that this was metaphorical, that neither viruses nor natural selection has any intent. Also, the readers of this site, I strongly suspect, all already know that.
“You should do that even if you’ve been vaccinated, too, as we don’t yet know whether vaccinated people, themselves resistant to illness, could still get a form of asymptomatic infection that could be passed on to others through, say, expelling viruses lodged in your respiratory tract.”
I’ve read this warning in other places as well, and it worries me because of the example of Marek’s disease in chicken (see the “prevention” section in https://en.wikipedia.org/wiki/Marek%27s_disease).
Could Jerry perhaps elaborate on how evolutionary pressures shift under the hypothesis of a vaccine that does not prevent infection?
You can link directly to a section of a Wikipedia article by adding a hash sign and the section’s name to the end of the article’s own web address: https://en.m.wikipedia.org/wiki/Marek%27s_disease#Prevention
(Note that if the section name is made up of more than one word they need to be joined using underscores _ . This doesn’t work with subsections within top level sections.)
Of course the biggest problem going right now is simply getting people vaccinated. That most import aspect is simply still not happening as it must to get us clear of this monster. I am speaking of the U.S. and more specifically some of the states that are doing the poorest job. Both Illinois and Kansas are down there at the bottom of the group of 50. Yet West Virginia is at the top and doing very well. This only shows that leaving it up to the states all this time has been a disaster and national support is necessary if we are going to accomplish the goal.
In the mix of this news are measures aimed at relaxing restrictions, now that we’ve seen a dip in infection rates. Here and there, restaurants are opening (but with social distance measures), and there is mounting pressure to open schools and allow young people to once again become active in sports.
Fascinating and very informative article. Thanks for posting. I’m waiting for Sunday now.
Greg, could you expand on this?
“New variants…seem to be increasing in frequency, and it’s apparently not due to random genetic drift because the increase is observed in several places. These parallel changes in frequency represent a good sign of positive selection.”
The way I understand the models of positive selection on DNA sequences, it’s not the geography that indicates positive selection but the observation of two (or more) versions of the same mutation increasing in frequency in different lineages of the virus phylogeny. Those parallel mutations could be in the same geographic location or different locations. But if the new mutation occurred only once (in one location), and then moved to another population (in a second location), and increased in frequency in both places, that wouldn’t be evidence of positive selection.
Is that what you were getting at – parallel mutations in different parts of the phylogeny?
Thanks, great post, very helpful.
I’m Jerry and wrote the post above, so presumably this goes to me. Yes, if the new mutation is a single base changes that spreads in different lineages independently, that is evidence for selection (it’s nearly the same thing, as geographically independent strains are independent lineages, too). But you’re incorrect in ruling out selection if a new strain increases in frequency in different places independently. If that happens three or more times, say, that is definitely evidence for selection if you can pinpoint the genetic differences associated with the increase. There are really only two explanations, drift or selection, and drift can’t explain parallel increases.
Woops sorry I misunderstood the Note at the top.
I have now edited the Note slightly to make it clearer.
GCM
Is it possible that an increased frequency in a particular variant has more to do with “super-spreader” events than selection? In other words, could an increased frequency of a variant which happens to spread at one of these events be confused with selection? Are there tests to distinguish them?
It’s possible, but if the increase happens in parallel in different places it would seem unlikely that the same variant is associated with superspreader events in different places.
To paraphrase the noted statistician Auric Goldfinger: Once is happenstance. Twice is coincidence. Three times is positive selection.
Perhaps ‘evolution’ is being mentioned because otherwise God is sending a plague and changing it to defeat medicine?
Some what off topic, but when I do the math, it seems that deaths from coronavirus are going to drop tremendously in the US soon.
1) Over 80% of the people who die are over 65.
2) The elderly are getting vaccinated first.
3) Once you get the second dose, you do not die from coronavirus. The NYT said ZERO out of 17,000 in the vaccine trials. In a podcast, Ben Shapiro said that no one in Israel has severe symptoms out of over 400,000 people to receive the second dose.
I expect daily deaths in the US to be down 75% (under 900) by the end of March and down 90% (under 300) by the of June. If you throw in cases already dropping, R falling from vaccinations and warmer weather, these actually seem pessimistic to me.
This assumes that none of the variants will have significantly different outcomes after the vaccine.
And the elderly population – or indeed any in the population with conditions that increase the likelihood of death – has decreased in number so much, the virus is running out of individuals to infect that will likely die.
IOW – the virus is running out of people.
We now have at least 3 distinct variants (mutations) that spread rapidly through increased transmissibility — Britain, S. Africa, & Amazonian Brazil. If the mode of advantage in each variant is different, say increased resistance to desiccation or heat, larger viral load, etc., does the Covid-19 virus exhibit recombination that would permit different mutations to join up in a single genome to produce a super-spreading strain? Perhaps someone who knows more about viruses than I do could comment.
Also, I find many discussions of the virus very short-sighted by not recognizing that there are many ways for mutations and selection to make the virus much much more dangerous for humans. Two possibilities that are particularly worrisome are 1) a new mutation arising that permits Covid-19 to infect and kill younger people (perhaps first taking hold in schools or multigenerational homes) and 2) mutated viruses that provoke unbearable pain in victims and increase their spread through the empathy shown by comforters.
I think everyone should consider themselves on lockdown until the problem of spread is solved.
Greg’ll probably have a lot more about this on Sunday, but in the meantime there’s a summary of five of the variants at https://www.scientificamerican.com/article/the-most-worrying-mutations-in-five-emerging-coronavirus-variants/. One of those it lists, from Spain, appears to have spread just by drift (/founder effect) because it happened to be common in Spain when the European travel restrictions were loosened last summer.
The other four variants, which appear to be spreading due to selection, share many of the same functionally significant mutations. The UK, South Africa, and Brazil P.1 variants all share the same N501Y mutation (that is, position 501 of the spike protein is changed from asparagiNe to tYrosine). The South Africa and Brazil P1. and P.2 variants have the E484K mutation, and the South Africa and Brazil P.1 variants have K417N or K417T. Each of the variants also has other non-shared mutations, but as far as I know these aren’t thought to be functionally significant.
So… there are a bunch of variants, but they’re all mutated in pretty much the same direction (so to speak). This suggests that a single modification to the vaccines (at least, the mRNA-based ones, which are easy to tweak) should boost efficacy against all of them.
I agree about lockdown, though. At least, as much as people are able to; not everyone has the luxury of being able to hide at home like I’ve been doing.
So several identical and independent mutations have been selected in different virulent variants. I’m a little surprised at this, but it shouldn’t be taken to mean that there is only one or a very few ways for spontaneous mutations and selection pathways to make the virus meaner. Thanks for the info and reference. W.
The elderly are not getting vaccinated first in my state. I do not know about other states. I spent most of last week pointing out to various elected officials that the system in place is unworkable. Seniors in my state have had to fight a pretty fierce uphill battle to be prioritized at all. As often happens the squeaky wheel gets the grease. I have an appointment to get my first dose tomorrow. I’ll have to assuage my guilty conscience by continuing to be a squeaky wheel. I also want to donate to efforts to get vaccinations to parts of the world that have problems getting any vaccine at all. I’m not paying for the vaccine so it makes sense to me to donate what I saved.
“Evolution deniers have no explanation for these new strains of Covid-19 that are spreading so rapidly.” Of course we have an explanation. God created the Covid-19 variants as punishment for the world-wide increase in
secularism. QED. Alternatively, the new variants were cooked up by Hillary, Nancy Pelosi, George Soros, Hunter Biden, and the Rothschild family in their secret underground lab under the Comet Ping Pong pizzeria.
Thanks for doing this jerry and for coordinating with greg for sunday. I liked your analogy to evolution of antibiotic-resistant bacteria. My guess is that many people can relate to that as many doctors have become more reluctant to prescribe antibiotics in recent years. We get good information from clinicians, virologists and epidemiologists in the mass media; but you are the first evolutionary biologist i have seen jump in. Cheers
Thanks for a really informative and helpful article. Latest news from the UK is that vaccination cuts transmission as well as protecting individuals.
Going for my first jab on Monday!
For anyone interested in what’s going on at the molecular level at the initial infection stage and via the latest ThisWeekinVirology, but mentioned sorta obscurely at the end, a 3D computer-graphics illustration of SARS-Cov-2 docking with a human cell and fusing with the host cell membrane, from Janet Iwasa’s lab.
What you want is what’s under the heading The Entry Animation
The spike protein is cyan, and once it gets to the membrane the “receptor”, three of our angiotensin-converting-enzymes – the purple dimers – come together to bind with the trimer of spike proteins. Then orange and yellow proteases (enzymes that clip protein chains) gather around and find sites that they can bind to (many, like these ones, are very specific for what place(s) in protein chains that they can clip) and clip sites on the spike protein. That’s what the little white flashes represent. Then the clipped pieces fall away revealing the part of the spike that cascades down thermodynamically and integrates into the cell membrane so the virus can dump its RNA payload into the cell.
That animation is remarkable.
Being a hypochondriac, I am worried about this virus evolution. If the virus is evolving to evade the natural antibodies, what’s preventing it from being able to evade the vaccine-induced antibodies? As you mention, this is what the antibiotic-resistant bacteria do and doctors are now more reluctant to prescribe antibiotics but with the vaccine aren’t we basically jabbing everyone and putting a lot of selective pressure on the virus to evade? Is it going to turn into a race against the virus?
The vaccine antibodies are “designed”, though the use of the designed mRNA that stimulates them, to be wide-ranging and cover lots of possible mutations. They’re also designed to prevent the virus from evading the antibodies to the vaccine as much as possible. Remember, natural antibodies act against more than just the spike protein, which is the sole bit of the virus for which the vaccines are designed.
Well, the mRNA is to the spike itself with just two substitutions for proline residues to lock the conformation of the protein into the pre-binding form to help give the immune system a leg up in recognizing the unbound form of the spike. (I vaguely recall hearing that there are other neutral mutations in the RNA, adjusting it to human codon preferences, but if that’s right it of course has no impact on the resultant amino acid sequence.) Also from somewhere I recall hearing recently that there are around 20 epitopes on the spike, so one or a few mutations shouldn’t disrupt general recognition of the spike protein by the immune system. Further, for a sense of scale, there are 1273 amino acid residues in the spike protein, so major disruptions starting from a single mutational event are going to be rare.
Is the spike protein glycosylated or otherwise modified?
Good question. Not that I am aware of. I’m currently trying to find out if any post-translation mods are known in any animal viruses. There is one example that I know of in bacteriophages (HK-97, specifically) of an isopeptide link formed by a Lys sidechain that adds to an adjacent Asn (with loss of NH3) in capsid assembly.) Here’s the link to the final paper specifically identifying that crosslink.
(But whenever I hear glycoslylation I’m reminded of long ago when one of my protein chem-oriented fellow grad students said, “Whenever I hear glycosylation my eyes glaze over – all I can think of is snot.”)
Great post Jerry, thanks. I have a nagging question re: selective pressures and Covid-19: As you are surely aware, there is increasing debate around the relative merits of strict adherence to the two-jab protocols used to assess the efficacy of Pfizer and Moderna RNA vaccines versus a “Just jab as many arms as you can as fast as you can, 2nd shots can wait” approach. What I have trouble getting my head around is whether the latter suboptimal strategy would, in fact, impose selection on the virus for variants better able to escape our immune responses– akin to the detrimental selective effects of half-finished antibiotic prescriptions. Is that analogy apt?
It’s a very good question, and that depends on the level of protection afforded by only a single jab. If that’s 95% as well, or close to it, then there’s not much less selection on the virus. I in fact don’t know why they decided to use two jabs for Pfizer and Modern instead of one. It depends on the total efficacy of one jab versus two, and I don’t think we know that.
Start at 40:00 in the most recent TWiV, that just dropped yesterday. Discussion of the Pfizer results just published in the New England J Med. At ~51:00, the graph of COVID cases in vaccine recipients v placebo. Well before the second jab @ 21d, the cases in vaccine recipients depart from controls, and essentially stay there. I don’t think it can be said that if there hadn’t been a second jab the line would be the same, but there’s no indication of a further decrease in slope at, say, 28d, as there might if added protection was kicking in a week after the second dose.
Lacking much in the way of conventional credentials in immunology and vaccinology, my guess is that the second dose is in line with the second dose of Shingrix, that it is expected to provide enhanced long-term protection, but for the shorter term this would seem to indicate that you’re already well on the road to protection after a week or so.
Hemp– Thanks SO much for the link to the TWiV podcast and the relevant clip! The data discussed by the well-informed panel does indeed suggest strongly that a single stick offers very close to the maximum measured efficacy observed following the full two course regimen, thus mooting my concerns re: selective pressure for viral escapes that may be cultured through a single-dose approach. Now if I could only get a jab!
I watch Bill Maher every week…last week, he had a couple evolutionary biologists talks about Covid origins. From “the lab-manmade” perspective…was an interesting talk. Nothing on this thread mentions it. Is that crazy talk? Maher is a sort of anti=vaxer…don’t quite know where he’s at. Just asking.
I didn’t see Maher, but thoughts re. SARS-CoV-2 having originated in a lab:
1) I remember hearing somewhere nearly a year ago from people tapped into the detailed interactions between the Spike and ACE-2, that if someone had designed the Spike for better interaction they wouldn’t have chosen the residues that are present.
2) The Chinese released the genomic sequence of SARS-CoV-2 almost as soon as it was possible to have determined it. All of the vaccine players took that information and ran with it, without which we’d be behind in this effort by the amount of time that elapsed between its emergence in Wuhan and whenever it appeared outside China where it would then have been possible to culture up the virus and sequence it.
“if someone had designed the Spike for better interaction they wouldn’t have chosen the residues that are present.”
A person could pick residues as an exercise, but if they wanted to get somewhere they’d use directed evolution because it works for such things, and can produce surprising results.
Brett Weinstein and Heather Heying were the guests.
Thx, I’ll look for it.
Or a team might use a different selection method, or combination, or even add additional things but my point is simply that one person, let alone a team, would never stare at sequence alignments then cogitate up residues to put in to make the virus exert some native effect with enhanced abilities because we know that e.g. directed evolution or artificial selection is easier and more efficient for such things, if that would even be feasible at all for this. Also, the results are sometimes surprising. They (meaning, a team, or a lab technician) might design expression constructs or some such for a technical necessity. But that might be evident in the sequence.
What I wrote has nothing to do necessarily with what Weinstein or Heying said – I was just saying who it was because someone wondered.
I think my “If someone had designed it…” take-home was based not on someone looking at sequence alignments but on knowledge of what residues in earlier SARS spikes are responsible for docking with the ACE-2 protein, and how those interactions might be improved if someone were to try to do that.
I’d include a link but it’s so long ago (probably all of 9mos!) and I don’t remember any good search terms so I don’t think I’d be able to find it again.
Oh that, yeah, sure, I see that as a necessary part of the process – find areas to direct evolution to. But I don’t see a true redesigning over the whole thing being done without a technique using selection, directed evolution, or something.
Interesting. Thanks for your inputs.
Perhaps a bit of a cheeky comment but wouldn’t the appearance of a new variant have to involve a change in the frequency of variants?
But more seriously:
I would be more cautious. Granted, I’m not an evolutionary biologist, but the fact that they are more spreadable means they have more opportunities to mutate, no? And given the sluggishness with which some countries are deploying or even obtaining their vaccines, that seems like an opportunity for even more disaster.
From what I’ve gathered from watching/listening to virologists closely engaged in the whole story, I gather that the verdict is still out regarding spread-ability of the various variants. There’s also the aspect that it emerged in the UK, which with its pub culture fosters COVID spread. One of Daniel Griffin’s podcasts on This Week in Virology in the past month touched on that.
It seems to me that the crucial thing will be to look at the fatality rates of people who have the “original” version (whatever that is, given all the known mutations already identified) vs. the those infected by a variant with whatever the mutation of concern is. That involves sampling a lot of victims and doing a lot of genotyping. I don’t doubt that some lab(s) is/are trying to do that, but I’m not aware of any papers that have come out about that. There’s also the aspect that treatment approaches are improving and the fatality rate is decreasing, so a comparison ought properly to be done within the same timeframe. Otherwise, if the rate of the original version, sampled earlier, showed the same as the variant sampled later, it could mean that the variant was more lethal, but it’s lethality had been tamped down by better treatment.
Interesting article. I have two thoughts.
First, you often hear from creationists that mutations can’t lead to increases in information or improve an organism. I suppose the first part is technically true, but not when the mutation is considered in conjunction with selection. For the second part, the mutation that led to the British variant appears to have improved its ability to infect new people. Good news for people fighting creationists but bad news for the World population.
My second thought is that making humans very sick or killing them is a bad trait for a virus to have in evolutionary terms. If COVID19 just gave everybody flu symptoms for a few days and did nothing else, we wouldn’t be trying anything like as hard as we are to eradicate it. I wonder if there’s a general tendency for viruses to evolve to be less virulent over long periods. I remember Steve Jones giving a talk about the evolution of HIV claiming that this was the case, for HIV at least.