Drs. Sanford and Carter respond to PS Scientists

[Faizal_Ali]

Welcome to the group!

[Faizal_Ali]

@PDPrice

You’ve appealed to junk DNA to justify your claim that effectively neutral mutations are balanced in their effects. But you’ve also denied that effectively neutral mutations occur exclusively in junk DNA. How do we put these things together?

I don’t see the problem here.

Suppose an organism is born with 100 mutations that its parents did not possess.

Also suppose that 90% of its genome is junk.

That means, right from the outset, at least 90% of these mutations can be expected to be neutral, since they would be occurring in segments of the genome that have no effect on function (i.e. are junk).

How much higher than 90% this number will be will depend on the proportion of mutations to functional parts of the genome that are neutral, beneficial, or deleterious.

Pretty simple, really.

[dsterncardinale]

@chris_doesdna2018

Even if we assume fixed fitness effects–regardless of environment or population size–GE still does not establish that those fixed effects are inherently deleterious.

Yep, it’s just layers of assumptions that range from merely unrealistic to completely off the rails.

[glipsnort (Steve Schaffner)]

@PDPrice

It certainly does when the parameters are realistic. We can’t run an experiment over millions of years, so this is the best way we can “empirically” verify the claims of evolution.

This has already been addressed by others, but my main complaint was the claim that this was the first time the selection threshold had been determined quantitatively. Would you care to defend that claim?

@PDPrice

I still think you’re being misleading with this claim. As addressed in the article, we certainly can learn valuable information about the effects of mutations–the overall spectrum–and that’s mostly consistent of effectively neutral ones, by everybody’s account, including yours.

I don’t know what you’re trying to say here. If you have evidence that most mutations of very small effect are deleterious, present it. You’ll be the first to do so.

[Michael_Okoko]

@chris_doesdna2018

GE assumes that the effects are both immutable and negative for the organism. The former claim is demonstrably falsified through data. I would argue that the latter is also demonstrably falsified through large-scale variant databases like gnomAD.

Both are falsified by the Lenski LTEE data too, I think.

[Rumraket (Mikkel R.)]

@Faizal_Ali

I don’t see the problem here.

Suppose an organism is born with 100 mutations that its parents did not possess.

Also suppose that 90% of its genome is junk.

That means, right from the outset, at least 90% of these mutations can be expected to be neutral, since they would be occurring in segments of the genome that have no effect on function (i.e. are junk).

How much higher than 90% this number will be will depend on the proportion of mutations to functional parts of the genome that are neutral, beneficial, or deleterious.

Pretty simple, really.

Technically mutations in junk do have small fitnesses associated with them, they are not strictly neutral.

But here @glipsnort’s analogy with using toner to print letters on pages shows why, at least with respect to junk DNA, the mutations that occur in junk DNA will have an equal distribution of deleterious and beneficials. That’s because to print every one of those letters carries a very small cost in terms of toner (just like copying useless DNA basepairs costs a tiny bit of energy). But this cost is very small, basically invisible to large multicellular eukaryotes with small relative population sizes.

Now, the different letters don’t have equal costs. For example it might cost more toner to print the letter M than the letter I, there’s simply more lines to print. A page consisting entirely of M will cost more toner than a page consisting entirely of I. So if you can only mutate between M and I and mutations are essentially random, at some point you’re going to reach some number of Ms in your junk sequence that reversal to I has become as likely as turning any remaining I’s into Ms too, and then you’ve reached an equilibrium where the distribution of deleterious to beneficial mutations in junk DNA is equal. That’s regardless of whether mutations are chemically biased more towards turning I into M, than M into I. At some point you’re still going to hit some equilibrium where you’ve got so many M’s that the probability of reversing to I equals the probability that any remaining I’s mutate into M. And then your distribution of fitness effects will be equally deleterious to beneficial.

[chris_doesdna2018 (Chris Hughes)]

Both are falsified by the Lenski LTEE data too, I think.

Yes, I would agree that Lenski’s LTEE directly falsifies both premises. However, microbial mutational accumulation assays are such strong evidence against the GE hypothesis that more recent interlocutors have excluded bacteria from its purview. I would suggest this is a case of special pleading. Rather than arguing over logical fallacies, I chalk it up to a silently acknowledged failure of the hypothesis to survive experimental scrutiny.

[Rumraket (Mikkel R.)]

One of the biggest problems with the whole GE concept is the completely bonkers(effectively physically impossible) assumption that the DFE of mutations must remain essentially constant no matter how highly adapted an organism is. That simply does not make physical sense. I dare say, it has to be physically impossible for this to be the case.

Living organisms are entities of atoms and molecules that function according to the laws of physics. They have weight, their constituents are held together by inter and intramolecular forces of attraction, and they must spend energy to perform their life-essential functions. This implies there are some absolute limits for how effective any given function in an organism can become.

At some point it will not be physically possible to improve some aspect of an organism in a way that contributes to it’s reproductive fitness, without that either compromising some other aspect of it’s physiology, or hitting some hard, fundamental physical limit.
There is going to be some relationship between the organisms total weight due to it’s mass, and it’s energy expenditure required to maintain this total mass. There will be some global optimum where it physically cannot grow bigger without the material of which it is made collapsing, or it will require more energy to maintain it’s body than it can ever hope to collect and consume.

At that point, this aspect of the organism’s physiology that contributes to it’s fitness cannot be further improved. That means there are no more beneficial mutations that can contribute to this part of the organism. That means the distribution of fitness effects of mutations, when at the top of the fitness peak, has been skewered entirely to neutral and deleterious. Either mutations are strictly neutral, or they are deleterious.

The most obvious example I can think of is enzymes. There is an absolute limit to how effective an enzyme can become where it cannot be further improved. This limit is defined by how fast the substrate can diffuse into the active site. It does not matter that it might be possible to make the enzyme able to pull electrons off the substrate even faster, because it would still have to wait for a new substrate molecule to diffuse into the active site. That means when a large fraction of the enzymes in your genome are approaching global, or even local optima, the DFE of mutations for them will start to skew more towards neutral and deleterious. The corollary of this is that, when you move further down towards lower fitness, many more beneficial mutations become available that were not available at or close to the top.

This explains why it cannot be true, in terms of absolute numbers, that the DFE of mutations is constant. There cannot be a fixed relationship between beneficial and deleterious mutations. Their relative numbers must change and reflect the level of fitness for the organism. Highly adapted organisms must have many more deleterious mutations available to them than very poorly adapted organisms.

Another related factor that also affects the DFE of mutations, and shows that it cannot be constant, is the phenomenon of diminishing returns espistasis. As I wrote in the DFE of mutations thread:

In reading one of the papers from the LTEE the concept of diminishing returns epistasis was mentioned, which also helps explain why it has to be physically true that the shape of the curve for the DFE of mutations must change over time as organisms become either more less well adapted.

Imagine an organism A that has some reproductive rate, and then a beneficial mutation occurs that increases it’s expected number of offspring by one. Now if this organism normally is expected to produce 100 offspring, and it has one more, it has improved it’s number of offspring by 1%.

Now imagine another organism B that has twice the reproductive rate that A had, and then this organism suffers the same mutation A did. It produces 200 offspring normally, and now can add one more. But now the mutation only has a 0.5% effect of improvement.

But that would imply the selection coefficient for this mutations has decreased. As the organism has gotten more fit, the effects of individual mutations have gotten smaller in proportion.

An analogy is that you are to push a heavy car that has run out of fuel to the nearest gas station. When you do it alone it’s very hard, but if you get another person to help, it’s half as hard, but then if you get one more, you’re doing a 3rd the work, and then with a fourth, a quarter the work. The gain from every additional person helping push the car becomes smaller and smaller, to the point of being neglible. When you’re 30 people pushing the car, you probably can’t even feel if any single person decides to take a break.

Again it is trivial to think of real examples where real biochemistry would result in mutations exhibiting diminishing returns epistasis due to the fitness-level of the organism. Where gaining two more ATP molecules per second means much less to an organism that can generate 1 billion per second than it does to an organism that can generate 1 million per second. Which means the same mutation that adds two more ATP molecules per second, can have a literally thousand-fold smaller effect for the highly adapted organism, than for the poorly adapted organism. Which means that the DFE of mutations has shifted, it can’t be constant.

So they can’t be constant in terms of absolute proportions benefical:deleterious, and they can’t be constant in terms of their magnitudes.

[chris_doesdna2018 (Chris Hughes)]

@Rumraket

Technically mutations in junk do have small fitnesses associated with them, they are not strictly neutral.

Technically they may have fitness associations, but I’m skeptical of claiming those associations outnumber strictly neutral sites on average. I’m curious if you have a citation or two to back this up?

[PDPrice (Paul Price)]

@glipsnort

I don’t know what you’re trying to say here. If you have evidence that most mutations of very small effect are deleterious, present it. You’ll be the first to do so.

There’s nothing more I can say to this; Drs Carter and Sanford addressed this gaslighting (?) in the article itself. All the data we have on mutations strongly and clearly show a massive skew toward deleterious. That’s why:

“In summary, the vast majority of mutations are deleterious. This is one of the most well-established principles of evolutionary genetics, supported by both molecular and quantitative-genetic data.”
Keightley P.D. and Lynch, M., Toward a realistic model of mutations affecting fitness, Evolution 57 (3):683–5, 2003

Consider the dead horse beaten.

@Rumraket

But here @glipsnort’s analogy with using toner to print letters on pages shows why, at least with respect to junk DNA, the mutations that occur in junk DNA will have an equal distribution of deleterious and beneficials.

I’m not going to debate Junk DNA; it was handled in the article. But even if I grant junk DNA, it doesn’t solve your problem. Dr Schaffner has already stated that he is not arguing that all effectively neutral mutations happen in junk DNA. That fact alone means the distribution can’t be even; there will certainly be more damaging than beneficial mutations in the spectrum. And that is not something that can be handled by evolutionary theory–it is fatal, both literally and metaphorically. I also happen to think Dr Schaffner is both educated and intelligent enough that he already knows this fact. Maybe he’s holding out for future discoveries to come to the rescue?

@Rumraket

At some point it will not be physically possible to improve some aspect of an organism in a way that contributes to it’s reproductive fitness, without that either compromising some other aspect of it’s physiology, or hitting some hard, fundamental physical limit.

That silly claim has already been addressed and debated on this forum. No need to repeat myself, other than to simply say again that there is no limit to how engineering solutions can be creatively applied to improve things. If you can’t get any faster in a car you can invent an airplane. And so forth.

@Michael_Okoko

Yeah he should answer it, because if GE leads to the eventual extinction of life forms, then it makes no sense that life has persevered for billions of years.

It hasn’t.

@structureoftruth

I think his answer to that would be that genetic entropy is evidence against the idea that life has persisted for billions of years.

There’s a right answer.

@structureoftruth

Of course, that just takes the discussion back to all the solid evidence we do have for the antiquity of life, earth, and the universe…

Not if we’re going to stay on topic it doesn’t.

[Rumraket (Mikkel R.)]

@PDPrice

That silly claim has already been addressed and debated on this forum.

It’s been addressed in the sense that you wrote an answer, and it was silly when you wrote it. I responded back by explaining how silly your answer was, and I’ll be happy to repeat that here.

@PDPrice

No need to repeat myself, other than to simply say again that there is no limit to how engineering solutions can be creatively applied to improve things.

Of course there is. The laws of physics themselves will at some point step in and prevent improvement. Take the enzyme again, the substrate’s rate of diffusion is fundamentally limited by it’s behavior in solution, which owes to the nature of the fundamental forces of attraction and repulsion that govern it’s atomic and subatomic constituents. It just can’t get better.

So how do you get substrate to the active site faster? Maybe you get a transport protein to put it in there faster than diffusion, but wait, your transport protein needs energy to catalyze the translocation of substrate, it needs to convert ATP to ADP. So now you need ATP to diffuse into the transport protein’s active site, which is limited by the rate of diffusion of ATP. So now you want to shuttle ATP to the transport protein with another transport protein… and on and on. You can’t get around it.

Some given protein that binds to DNA just can’t bind any better without taking up too much space. There’s a physical limit to how strong chemical bonds can get, and of course if they get too strong they can’t be removed again. Etc.

@PDPrice

If you can’t get any faster in a car you can invent an airplane. And so forth.

Congratulations, you’ve discovered the concept of niche construction, which is to change the fitness landscape of your environment by your own evolution, and adapt to a different mode of life, not to actually evolve higher reproductive fitness.

But that too would entail a fundamental change in the distribution of fitness effects of mutations through historical contingency and conditional sign epistasis, because now mutations that used to negatively affect your ability to drive fast, such as big strong wheels, might positively benefit your ability to land or take off. An entirely new spectrum of mutations has to be established, the old one no longer applies.

[Rumraket (Mikkel R.)]

@PDPrice

I’m not going to debate Junk DNA; it was handled in the article. But even if I grant junk DNA, it doesn’t solve your problem.

It actually goes a long way towards answering why many mutations would have very small effects, as since the genome is mostly junk most mutations will occur in junk, and junk-mutations mostly have small effects. So this would partially contribute to the height of the curve near neutral.

And just to stave off any confusion, I’m not saying the junk protects against deleterious mutations in functional regions as the rate of mutation does not depend on the amount of DNA.

@PDPrice

Dr Schaffner has already stated that he is not arguing that all effectively neutral mutations happen in junk DNA.

Neither did I. You seem confused.

@PDPrice

That fact alone means the distribution can’t be even

No, that fact does not mean that at all.

@PDPrice

there will certainly be more damaging than beneficial mutations in the spectrum.

Across the genome in total? I would not dispute that, but I would dispute the claim that the distribution is constant, and that it will always have an insurmountable fraction in the zone of no selection. That claim is physically unrealistic, and is in fact contradicted by experiment.

@PDPrice

And that is not something that can be handled by evolutionary theory–it is fatal, both literally and metaphorically. I also happen to think Dr Schaffner is both educated and intelligent enough that he already knows this fact. Maybe he’s holding out for future discoveries to come to the rescue?

Now of all silly things said in this thread, that one has to take the cake.

[John_Harshman]

@dsterncardinale

I assure I’m well versed, but thank you for the gratuitous comment.

What you describe is not Sanford’s model as I understand it. Are you sure I’m wrong?

@dsterncardinale

Sanford assumes that fitness within populations will be equal enough so as to preclude selection.

No, he assumes that the component of fitness contributed by nearly neutral alleles will be equal enough so as to preclude selection. I don’t think he doubts purifying selection, i.e. that there are alleles that are individually deleterious enough to be subject to selection. Are you sure? GE is about the accumulation of nearly neutral alleles, not the accumulation of significantly deleterious alleles.

@dsterncardinale

If you think otherwise, that Sanford’s premise is valid, then by all means, what data do we have that support it?

What data could there be? If the selection coefficient is low enough that the allele is nearly neutral, there’s no way to detect it. Selection coefficients that that can’t be detected in the lab are enough to be selectable in wild populations, and nearly neutral alleles are even less detectable.

[Rumraket (Mikkel R.)]

@chris_doesdna2018

Technically they may have fitness associations, but I’m skeptical of claiming those associations outnumber strictly neutral sites on average. I’m curious if you have a citation or two to back this up?

Afaik the different DNA basepairs A-T and G-C don’t have exactly equal metabolic costs, among other things because they don’t have exactly equal lengths of pathway biosynthesis. I don’t know the actual exact differences, but in so far as they are not exactly equal, they can’t have exactly equal fitness costs of replication unless there is some other indirect physicochemical effect they have that somehow exactly balances out their cost of biosynthesis, which just seems doubtful to me.

[dsterncardinale]

@John_Harshman

Are you sure I’m wrong?

Yes. We can go chapter and verse if you want through the early chapters of “Genetic Entropy and the Mystery of the Genome”.

@John_Harshman

@dsterncardinale

Sanford assumes that fitness within populations will be equal enough so as to preclude selection.

No, he assumes that the component of fitness contributed by nearly neutral alleles will be equal enough so as to preclude selection. I don’t think he doubts purifying selection, i.e. that there are alleles that are individually deleterious enough to be subject to selection. Are you sure? GE is about the accumulation of nearly neutral alleles, not the accumulation of significantly deleterious alleles.

If it’s causing enough problems to cause extinction, then it isn’t “nearly neutral”. It’s significantly deleterious. Which brings us back to the question of relative fitness within a population. If there are differences (which you seem to agree there are), then the least bad genomes will be selected for. Do you dispute that?

[PDPrice (Paul Price)]

@John_Harshman

What you describe is not Sanford’s model as I understand it. Are you sure I’m wrong?

Based on what I see here, you certainly do understand Dr Sanford’s thesis better than Stern Cardinale does. I’ve explained it to him countless times, but he is one of those who will not understand. And certainly not if it is coming from me. Maybe you can explain it.

@John_Harshman

What data could there be? If the selection coefficient is low enough that the allele is nearly neutral, there’s no way to detect it.

Perhaps not in isolation. But as mentioned in the joint article, we can look at the overall effects of mutations as a whole as we do mutation accumulation experiments. Most mutations are nearly neutral. So to say we have “no idea” what the DFE is for nearly neutrals is just wrong. We know it must be overwhelmingly negative, just as most mutations are overwhelmingly negative. This is both data-driven and intuitive. It is much easier to break a functional machine with unplanned changes than it is to make improvements on it. This is so obvious it should not need to be debated by intelligent people.

@dsterncardinale

If it’s causing enough problems to cause extinction, then it isn’t “nearly neutral”.

Yes, it is. The extinction only comes from the cumulative load.

@dsterncardinale

It’s significantly deleterious.

Only in combination, as a total deleterious mutational load.

@dsterncardinale

If there are differences (which you seem to agree there are), then the least bad genomes will be selected for.

Every member of the population will have inherited a new set of mutations, and nearly all, if not all, of these will be detrimental. Nearly all of them, if not all of them, will be effectively neutral. That will still be the case, even when you pick the “least bad”. But since the differences are so tiny, it doesn’t matter anyway. Other non-genetic factors (noise) will play a much stronger role in survival.

[chris_doesdna2018 (Chris Hughes)]

@Rumraket

Afaik the different DNA basepairs A-T and G-C don’t have exactly equal metabolic costs, among other things because they don’t have exactly equal lengths of pathway biosynthesis.

I think that’s an interesting perspective, but I’m not quite sure the cost-benefit analysis is straightforward for single-nucleotide variants. In either transition or transversion, the ratio of purine to pyrimidine should remain the same. It’s been a while since I’ve thought about nucleotide synthesis, but each synthesis step is regulated by the availability of ATP/GTP or CTP–it’s not clear to me why there would be overall differences in metabolic cost. Guess I’ll go look at the pathway