So…yes? Each intrahost population is a separate population, genetically, from all the other intrahost populations? Because that was the question.
2 Likes
Every single transmission is mutated?
1 Like
I’m not seeing that a strong bottleneck would produce a small probability of the best genotypes being transmitted. Would you mind elaborating?
1 Like
Why on Earth would we assume such a thing?
It is entirely possible selection maintains mean viral fitness of the population near some peak in the fitness landscape even if the most fit variant goes into a slight decline every time it re-emerges, as once fitness gets low enough the magnitude of gain from the most fit variant will push it right back up. Here’s why:
First of all if there is a 10% chance that the most fit mutant is transmitted, then on average 1 in 10 transmissions contain the most fit mutant. 1-2-4-8(there’s likely to have been one already as 15 total transmissions have occurred now)-16 (we’d expect at least one highest-fit transmission more now) -32(likely three more)-64(probably about 6 more highest-fit transmissions now) and so on. More and more most-fit mutants will crop up as the number of total infected increases.
Second you have completely ignored the effect of selection on transmissibility. If the most fit mutant is 20, 40, or 60% more transmissive, for example, you need to multiply it’s transmissibility accordingly(and reduce the infectivity of less infectious mutants). More fit variants infect more than less fit variants. The further down in the fitness landscape we go, the bigger the possible effects of mutations become. When the population is very close to the peak in the landscape, the gain from most fit variants might be very low, but as the population declines the possible gains increase.
Of course when a new pandemic starts, it doesn’t start with the most fit variant already. The virus usually has to adapt to it’s new host, so it’s definitely not at the peak of the landscape. Huge gains in fitness are possible, and that’s of course what we see happening around us today.
None of these factors figure in your calculation. You just assume that there’s a mean fitness decline with no effect of selection on infectivity for deleterious or beneficial mutants, and that the highest-fit or higher-fitness mutant never transmits. None of that makes sense. You haven’t so much shown or argued why there should be a fitness decline as you have simply assumed it by ignoring the effects of selection both between and within hosts.
3 Likes
Why? The best viral genotypes are those that reproduce the most efficiently and have the highest probability of being transmitted to the next host. Numbers you have invented do not make a convincing argument.
Your mean fitness reduction of 0.001 is anther number you simply invented. Also, you switched from talking about the maximum fitness initially to the mean fitness after 20 generations. And you neglected any possible increased fitness within any given host.
Please make an effort to produce arguments that take longer to dismantle than they do to read.
2 Likes
Yes, most probably
Here is how Sanford explains it:
When a virus is transmitted from one individual to the next, the first individual harbors a viral swarm. The second individual becomes infected by a random subset of that swarm (conceivably a single gentype). With this type of bottlenecking, the « best » viral genotypes within the first swarm have a small probability of being transmitted to the next host. This probability becomes especially small when infection arises from a single viral particule. Given a high mutation rate and regular bottlenecks, the operation of Muller’s Ratchet becomes quite certain, which should result in a continuous ratchet-like mutational degeneration of the viral genome.
It seem like all viral genotype should have roughly equal probability of transmission - unless there is some mutation that affect transmissibility, of course.
Is this true? I thought single virii are unlikely to cause infection, and that larger “doses” are usually required?
3 Likes
His explanation is wrong. That’s why he isn’t citing any data in it.
That’s utterly false for RNA viruses in multiple ways, starting with the fact that the first individual is subjecting the virus to selection, so the inoculum for the recipient is in no way random.
You might want to look up defective interfering particles. Their existence alone makes a mockery of Sanford’s explanation.
It is not true; you are correct, Dan.
The ID50 (infects 50% of subjects) for influenza contains hundreds to thousands of viral genomes. Because many of those are defective, this started with metrics like “particle-to-PFU ratio,” in which particles initially were counted by electron microscopy. Genomes are far easier to quantify by PCR now and seem to be the current standard for quantifying particles.
2 Likes
As far as Influenza is concerned, it seems that in most cases, 1 to 13 viral particules are transmitted from a human host to the next.
If what you claim here was true, strong bottlenecking would not be associated with fitness decline. But since it is well known that strong bottlenecking is indeed associated with fitness decline, your claim is false.
@Giltil quotes Sanford…
The second individual becomes infected by a random subset of that swarm
The alpha case infecting agent has a head start over downstream mutations. There may be gazillions of mutant viruses, but they are still swamped by the original genotype. For influenza, transmission is often pre-symptomatic, making transmission of the original copy more likely. As has been mentioned, then selection also plays a role. So Sanford is wrong to say that The second individual becomes infected by a random subset of that swarm.
If you go to a database like GISAID or the NCBI Influenza Virus Resource, and do a search without collapsing identical sequences, you will get more results. Given the small proportion of sampled specimens in reservoirs, it is reasonable that nucleotide identical infections have been transmitted by the thousands.
According to Worldometer, we are now up to about 130 million recorded cases of Covid, and the likely number of infections is probably much larger. The number of base pairs for this coronavirus is 30,000. Times four is 120,000. So by orders of magnitude it is arithmetically impossible that every transmission be mutated. Were Sanford even close to being right; well, even on the same planet, the Covid virus would have turned to puddles before it left Wuhan.
The aforementioned GISAID site is a serious research tool and fantastic visualization of the history of COVID. You can drag a time slider and see every mutation in the virus pop up as you move it. The empirical pace of mutation is nothing like the per transmission rate proposed by Sanford. As a bonus, the variants of concern have been identified in the literature, and adaptive mutations analyzed. I would invite experts on Genetic Entropy to identify the specific mutations which are deleterious or very slightly deleterious, and detail how they act as liabilities to the virus.
1 Like
It might help to distinguish between the particle count and the haplotype count.
It also might help to define “seems” in this case. Is Gil using it as a cover for cherry-picking?
The authors appear to be using a nonstandard definition of “particle.”
Agreed. It is difficult to express how spectacularly wrong that is.
I see you’ve made no attempt to justify the numbers you used. Nor does your supposed rebuttal even address my objection.
Why should the best viral genotypes have only a 10% chance of being transmitted to the next host?
1 Like
How can you appeal to bottlenecking or anything else to assert that fitness must decline, when it is evident that in the case of COVID, fitness has not declined but has in fact increased? It is evident that what you suggest did not happen, therefore something is wrong in your analysis.
1 Like
Yeah that suffers essentially all the same flaws I pointed out up in my previous post. Completely ignores the effects of within and between-host selection.
3 Likes
I see you’re missing the coded response, Roy. “But since it is well known that…” followed by no evidence should be replaced with, “But this one goes to 11.”
It’s a Particle, Man! 
Roughly 60% of SARS-CoV-2 transmissions have no mutations.
1 Like