If Sanford hasn’t fully developed his ideas in scientific papers, that is his prerogative, but not anyone else’s problem. If he expects the scientific community to buy his book first, then he has no basis for complaining that scientists are not engaging his ideas.
5 Likes
Okay, we’re done here, thank you.
1 Like
The difference in energy needed for replication would probably be swamped by the energy needed for transcription and translation, not to mention homeostasis and actual function in the environment. You only need to copy a ribosomal gene once, but transcribe it millions or even billions of times over the lifetime of an individual cell. Tighter control on leaky RNA transcription would probably negate the cost of many SNP’s.
Good day, sir. I said “Good day!”
You have apparently, in some way, proven your superiority to your own satisfaction. But does that aid any dialog?
3 Likes
The real question is how many of these nearly neutral mutations would need to accumulate before they would result in the extinction of a species. I don’t see where Sanford has ever answered this question.
1 Like
You will note that there is no reference cited to support that assertion, just “it seems unlikely”, with no argument as to why it should be unlikely and no citation to the literature.
No it isn’t. I don’t know why you should think so.
This is not a point at issue. Of course it’s true.
No, it has nothing to do with that. It’s that the mutation is so slightly deleterious that drift is stronger than selection; it could also be slightly beneficial or strictly neutral, and there’s no way to tell. Reproductive fitness is absolute fitness or relative fitness, depending on what you’re calculating.
1 Like
Look, I’m not interested in proving superiority. I am interested in not wasting my own time.
Here’s what I’m reading:
“I’m not trained in this field and I haven’t read the book on the model in question, and also I am 100% certain that your critique is flawed and that you don’t understand that model, and I don’t see how what you’re saying is relevant”
That’s fine. Believe what you want. Not worth my time to convince you otherwise.
Then you should respond to my criticisms with valid arguments. That woudn’t be a waste of time. Are you in fact a population geneticist?
But what is the amplitude in the reduction in fitness?
Boiling away in the background of this discussion is comparative genomics. If we align orthologous genes from humans, chickens, and fish what would we see? For many genes we would see strong conservation in exons and no conservation in the vast majority of intron sequence. This simple observation needs to be explained by Sanford’s model. If these small changes are so damaging then how can the introns be so different between these species while the exons are much more similar?
No, we see an overall picture of what mutations with detectable effects tend to do.
Nearly all mutations in a virus will be non-neutral because of the enormous effective population sizes.
If most mutations are nearly neutral, then that is a reason not to extrapolate their effects from the minority of non-neutral mutations.
Do you ever feel bad for saying things refuted by your own sources?
the fate of which is largely determined by random genetic drift.
Largely determined means that selection does play a role, but the role of drift is stronger. On one side of the threshold, selection is the dominant force, on the other side drift is the more dominant force. Both are always at play.
the rate of loss of fitness per generation may amount to 10^-7 per generation.
So for Hominini that’s a 10% reduction in 25Ma. One 0.01% beneficial mutation every thousand generations or so and we’re fine. Although realistically, I’d guess there is sufficient climatic variation over 25ka to account for a 0.01% difference in fitness anyway, so really you’d need… no immediately beneficial mutations to maintain parity. Funny what happens when you actually break out the math, eh?
1 Like
If it makes so much sense, why can’t you describe the thesis of the book in your own words?
Why don’t you read the population genetics literature. AFAIK, every quote you’ve offered here is second-hand.
Don’t you think that demanding that others read the book when you clearly don’t read the primary scientific literature is a bit asymmetric?
The fact that H1N1 is not extinct means that Sanford is wrong.
Even though the influenza paper is incoherent and misunderstands epidemiology, it may be suggested that is a microbial case, and GE still applies to higher organisms. One such candidate would be flies. They have a body plan, are suitably complex, and reproduce sexually. And reproduce they do, on the order of about two weeks into their life cycle. Given a creation date of around 4004 BC, that works out to greater than 150,000 generations, from Adam fly to the ones buzzing around now. From hiking, I can assure you that flies are far from extinct, and despite their long ancestry that deleterious mutations, strictly neutral, nearly neutral, or selectable, has not made a dent in their swarming population. How this readily available empirical fact squares with Sanford’s inexorable, cumulative mutational load and collapse does seem a bit problematic to his idea, ever if YEC were true and all of science false.
2 Likes
They showed it with regards to H1N1. With regards to humans, nobody knows. But we are certainly on that trajectory.
Alright, but that statement itself is from the literature, and it was made by people in a position to have an educated opinion about it.
You must have misunderstood me, because in other places I’ve seen you say the exact same thing. Something can be deleterious in terms of absolute fitness yet nonetheless unselectable with respect to reproductive fitness. That’s all I meant by that.
Certainly not as I’m defining the terms.
There is a way we can know general trends. That’s been addressed. As has your claim about strict neutrality, which the experts in this flatly reject.
Kimura was on the right track, but he greatly underestimated where the threshold of selection should be. His model unrealistically ignored the issue of the princess and the nucleotide paradox. Some changes are bound to be too small to be selectable, regardless of population size. And he also ignored noise, which is a huge factor. His model only took population size as the determiner of the threshold of selection. His formula was 1/2 Ne, or 1/2 effective population size.
See above. Kimura’s math was clearly unrealistic. But that’s not even the important part. The important part is that he understood that there would indeed be a net loss of fitness due to effectively neutral mutations. The only way around this is to invoke compensation by rare mega-beneficial mutations, which is what Kimura claimed. But he provided no evidence, and he didn’t so much as attempt to model it. In reality, this doesn’t make any sense at all. Why?
-
Fitness is not a substance. It’s a by-product of a functional genome, which is comprised of information. Information is not a binary substance that you simply “improve” or “damage”. It’s all contextual. Therefore it’s simply nonsense to suggest that occasional beneficial mutations could somehow outweigh the gradual corruptive force of mutations on the whole.
-
Linkage or hitchhiking (previously discussed).
-
Antagonistic epistasis. Much of what is deemed “beneficial” is only beneficial in a limited and reductive sense. This kind of “benefit” doesn’t work in combination with other examples of these “benefits”. It all comes crashing down.
(Responding to my own post since no one else seems to be interested in getting quantitative about this question.)
So in my simple model, how many VSDs are removed each generation because more highly loaded individuals produce fewer offspring? Using the numbers above, and doing a simulation because that’s easier than writing down the integral correctly, I find that ~1 VSD per generation should be removed by this effect. Since ~50 new VSDs can be assumed to arrive per generation, this does not seem like an effective brake on the accumulation of VSDs for a species like current H. sapiens. Unless my simulation is wrongly coded or represents a bad model, of course. But if it’s wrong, I’d like to understand how it’s wrong.
1 Like
I’m in the classroom full time now, but my training is in molecular and population genetics. My thesis work was on this exact question of mutation accumulation and fitness.
Here’s the deal. I asked about the equation for selection coefficient because the equation is s = 1 - W, where W is relative fitness.
W is the fitness of the thing your care about divided by the maximum fitness in the population.
So for there to be a selection coefficient > 0, there must necessarily be detectable differences in relative fitness. Detectable to selection, that is. If there are no such differences, if everyone in the population has the same relative fitness, than the selection coefficient on all of those different genotypes will be 0 (because everyone’s relative fitness will be 1).
So to say something like “there might be a selection coefficient greater than zero but it would be too small to affect relative fitness”, or anything along those lines, is a contradiction; a non-zero selection coefficient requires selectable differences in relative fitness. A selection coefficient is not an inherent characteristic of an allele or genotype; it’s a context-dependent product of the genotype, population, environment, etc.
And then bringing this back to Sanford, his model requires that relative fitness stays at exactly 1 across the board as absolute fitness declines (otherwise the different genotypes can be selected for, and that breaks his model), but then it all has to collapse to 0 all at once (otherwise, again, you have variation in relative fitness, different genotypes are selected for, model broken) for s to become 1 and the population goes extinct. This is what I meant above by “fitness cliff”.
I have read Sanford’s book very carefully, I’m quite well versed in the ins and outs of his model, and particularly the inherent underlying assertions. It’s worth reading, if for no other reason than seeing just how bad it is, and I recommend you read it before characterizing my understanding of it.
10 Likes
clearly, the idea of ‘fitness’ is actually divorced from natural selection/selectability in population genetics. This is the unspoken division of absolute vs. reproductive fitness
No it’s not. s = 1 - W where s is the selection coefficient and W is the genotype’s relative fitness. Sanford argues that selection would be too weak to counteract other forces like drift on slightly deleterious mutations. Even he’s not saying fitness and selection are unrelated because doing so is a crime against math
Also, I don’t know how you can claim absolute fitness is declining in humans when our population is growing. This one is just a crime against definitions
6 Likes
I wish critics would spend more time with this. The fundamental slight-of-hand for GE is co-opting the ability for non-neutral variants to act neutrally and possibly reach fixation in a population. However, GE ignores the population and environmental requirements for this to occur.
If the variant does not affect the fitness of the organism, then it is neutral. Period.
4 Likes
The compact size of viral genomes also means the majority of mutations occur in functional DNA.
7 Likes
No, they didn’t. Influenza viruses are still kicking along just fine.
6 Likes
Apparently you don’t understand how scientific literature works. Every statement must be supported either by evidence or by citation of a publication containing the evidence. Anything less can’t be used in support of any claim.
What you say is incoherent and not at all what I said. Are you perhaps saying “reproductive” when you mean “relative”? And when you say “something” are you referring to a large assemblage of alleles rather than just one?
You seem not to have defined the terms. What are you talking about?
What is this “way”? And you have exactly one expert rejecting strict neutrality, without any support, as I have already noted.
Not true. At the limit, an infinite population, there is no drift at all, and any least departure from strict neutrality is selectable. What is “noise” if not drift?
Works for me. That’s been my argument, though in purely qualitative form, all along.
Sorry. I thought you were asking for something more complicated.
Not true. Slightly deleterious mutations, invisible to selection because drift dominates, still have non-zero values of s. It’s just that those values can’t be determined. That’s why they’re called “slightly deleterious”. The whole idea of nearly neutral evolution depends on that.
Also, it’s not relative fitness that’s important for GE; it’s absolute fitness, which declines while relative fitnesses (of the accumulated slightly deleterious mutations) remain constant, and thus invisible to selection. See @glipsnort’s model, above.
I would claim that population size is not relevant to a selection coefficient except in a few odd cases.
Can you find an instance of such a claim for his model? That’s certainly not a requirement in order to GE to happen. All that’s needed is for the contribution to fitness of nearly neutral alleles not have significant variance in the population. Would you agree? If so, that does seem like a reasonable model to me. That’s why we have so much junk DNA, with no significant variation in genome size in the population. There is clearly some cost to having a large genome, but there’s equally clearly no selection against increasing genome size if there is no huge within-population variation. There also seems to be no significant effect even of very large genomes on absolute fitness, which suggests a problem with GE: if the distribution of slightly deleterious mutations is biased toward really, really slightly deleterious, even the massive accumulation proposed by GE still won’t cause a problem even over evolutionary time.
So much for nearly neutral theory. In order even to talk about nearly neutral evolution, it’s necessary to talk about selection coefficients too small to have any effect on fitness, given the population size. Note also that in order for fitness to be additive, it’s necessary that selection coefficients too small to have any effect be able to pile up into a measurable effect. Perhaps there’s a conflict between measurable fitness and mathematical fitness? But whether fitness is measurable depends on how you measure it. S too small to be determined by experiment can still result in fixation over evolutionary time in real populations. The larger the population, the smaller s can be measured. And so on.
1 Like