Agree. But in the case of flu, it would be difficult to argue for the adaptative value of evolution toward avirulence.
We do see an adaptive pattern away from virulence in the case of the flu. That’s why vector species are so important in understanding the emergence of new virulent strains.
1 Like
No, they don’t. I really invite you to read their paper to see that this is wrong. What they’ve done is to examine the genetic changes arising within H1N1 from 1918 to the present. And what they found is stated in their conclusion below:
While there have been numerous adaptations within the H1N1 genome, most of the genetic changes we document here appear to be non-adaptive, and much of the change appears to be degenerative. We suggest H1N1 has been undergoing natural genetic attenuation, and that significant attenuation may even occur during a single pandemic. This process may play a role in natural pandemic cessation and has apparently contributed to the exponential decline in mortality rates over time, as seen in all major human influenza strains. These findings may be relevant to the development of strategies for managing influenza pandemics and strain evolution.
To sum up:
- they document the numerous genetic changes that occurred in the H1N1 genome throughout its history.
- they provide evidences that most ( but not all) of these genetic changes not only are non adaptive but also likely degenerative.
- they note that the mutation accumulation is associated with the historical exponential decline in H1N1 human mortalities
- by combining 2) with 3), they (rightly IMO) suggest that the decrease in the bug virulence, far from being adaptive, is due to genetic entropy.
What are the evidences for your claim?
The Sanford & Carter’s paper presents strong evidences of the contrary, I.e., that since its emergence in 1918, the H1N1 genome has linearly accumulated non adaptive mutations with time and that this phenomenon is positively correlated with reduced death rates.
Even when positively correlated? How exactly are both apples and oranges in the context of viruses?
Carter and Sanford didn’t do this because it was impossible for them to do this.
Not really. See my answer to @swamidass below:
What is Viral Fitness and How is it Measured? - #24 by Giltil
Basically, their claim is that most (but not all) of the mutations that has accumulated in the H1N1 genome throughout its history are non adaptive and even degenerative. And contrary to what you say, they provide evidences for this claim. So no, it’s not the FIN, not at all😉!
Why? It seems to me this facilitates the long-term persistence of the virus in the human population.
2 Likes
I guess that human weight and height are positively correlated but they are not the same thing, right? The same apply to virulence and fitness. They often are correlated, but are not the same thing.
It’s defined the same way. The specifics that contribute to fitness differ, but the definition doesn’t. Absolute viral fitness (between hosts) is in fact just the basic reproduction number (R) that has become widely known during the pandemic.
Producing more offspring not using sexual reproduction. What’s the rat equivalent of producing more offspring using midwives, as humans do?
3 Likes
Virulence is tangential to the question of fitness, in my estimation. It is a bit like asking if increasing the ability to fly is a factor in fitness. It might be true in birds and bats, but isn’t true in most lineages. What we are really asking is how a specific viral lineage competes against other viral lineages in the same population, and virulence may have nothing to do with it.
2 Likes
The 1918 strain (we have the full genome) against one of the more recent, lower-fitness (according to C and S) descendant strains.
1 Like
What they’ve done is misrepresent the genetic changes. They are conflating reassortment with mutation.
If they truly believed that, they would be developing such strategies, but they aren’t. Why?
It misrepresents the evidence.
The graph isn’t linear and it conflates reassortment with mutation.
That is objectively false. Fatality rates vary from year to year with no obvious trend.
2 Likes
This is what I asked for.
LOL, ignore this part. I had something else in mind, but typed it another way.
Okay, let’s do this in a wee bit more detail.
C and S claimed two things indicate a loss of fitness over time from 1918-2009: Lower virulence and decrease in host-specificity of codon bias.
I’ll start with codon bias, because, coincidentally enough, literally half my PhD thesis was on the evolution of codon bias in viruses. For RNA viruses, you know how strong translational selection is? (That’s selection to match the codon profile of your host, btw.) It’s basically non-existent. Which means that codon usage in RNA viruses is basically random, and substituting one codon for another is, as far as we can tell, completely neutral. There are some exceptions around the edges: Humans don’t like CpG dinucleotides, so we tend to see a loss in those codons over time; those changes are adaptive in the virus. And guess what? C and S documented many such changes. So much for being entirely neutral or harmful. So that’s codon bias. To the extant any of the changes would be non-neutral, it appears they would be beneficial.
And then virulence. Oh, virulence. C and S commit the very basic mistake of assuming greater virulence = more fit. Ask an epidemiologist if that is a reasonable assumption.
Virulence correlates with fitness based on the ecological context in which a virus exists. If there are a TON of susceptible hosts readily available, then the limiting resource is cells to infect within each host. This is called intra-host competition. To be successful (i.e. for your genotype to propagate), you need to win the competition within your current host. That means infecting fast and making a lot of offspring as fast as possible. This will likely be very harmful to your host, but that doesn’t matter, because there are plenty more where that came from.
But what happens after like half the population has been infected and is now either dead or immune? Turns out, hosts can be a bit harder to come by. That means the rate-limiting step in your life cycle isn’t being the best within your current host anymore. Now the problem is getting to a new host. So what happens if you’re really good at competing in your current host, and therefore really deadly? Exactly. You kill your current host before any offspring can spread to a new host, and now your genotype is extinct. In this situation, it’s inter-host competition that matters; you’re competing with viruses in other people for access to the few remaining susceptible hosts. Under these ecological conditions, a better strategy is often lower virulence, less sick, less dead host, more opportunities to transmit.
This is exactly what happened with the 1918 pandemic. That strain did not vanish in 1919 or 1920. Nor did we suddenly get REALLY good at treating influenza and the secondary respiratory infections. But the mortality rate plummeted because the selective pressure on the virus flipped from intra-host competition to inter-host competition as the pool of susceptible hosts shrunk. And I will note for good measure that these changes are not just “breaking” stuff. Virulence is a phenotype that is very tightly regulated. The rate of viral replication and number of progeny are tightly regulated. (I once saw a really cool talk on this called “How A Virus Counts To 1000”.)
C and S consider precisely none of this. They just assume, for absolutely no good reason, that higher virulence = more fit.
So no, this is not an open question. They got basically everything wrong they could possibly have gotten wrong when characterizing the evolution of the 1918 pandemic influenza strain.
6 Likes
Cool points. I am definitely going to convert this page to PDF when it is closed. Its an excellent resource.
Good point. Thanks.
Key point. Thanks.
This is wrong, for they indicate 2 other evidences in addition to the 2 you mentioned, that is1) the extinction of all human influenza strains existing prior to the H1N1 strain; and 2) the apparent extinction of the human lineage of H1N1 in 1956, and then again apparently in 2009.
This is surprising. Do you have publications that support this point?
Then, why was there a codon bias in H1N1 in the first place?
Where did C and S document these changes? As for me, it has escaped my attention !
Huh? There are many influenza strains that did not go through extinct…
1 Like
First, the reason strains come and go is called strain replacement and it happens all the time, as a consequence of the dynamics I described in my longer post. If you’re well adapted for inter-host competition, what happens when a new strain to which very few people are immune appears? It outcompetes you within each host and becomes the dominant strain. But more importantly, there were and continue to be many strains that circulate at varying frequencies. Check the CDC advisories going back each year. C and S should have - they would have found their supposedly extinct strain in circulation.
Love when I get to cite one of my own papers.
There’s always codon bias. The thing is what’s driving it? Biased mutation rates, translational selection, enzymatic activity, biased repair mechanisms, etc. For RNA viruses, “translational selection” is not the answer, as C and S seem to think it should be.
Figure 7.
3 Likes