Mutation Rates Argument

[AC83]

Vox Day claims to have mathematically shown that "evolution is impossible". His argument appears to be that simple math shows that there isn't enough time to account for the genetic differences between humans and chimpanzees.

The argument looks badly flawed to me, but I am not a biologist and am likely missing or misunderstanding some things. My question is if I have correctly identified problems or if I am getting things wrong myself.

The JFG-VD Debate - Vox Popoli

We’re told that they [humans and Chimpanzees] are 98% identical in terms of DNA. So that would imply that there are sixty million different base pairs, which would tell us, because it’s two different paths, we’re talking each – they both have thirty million mutations since the common ancestor. And that’s kind of close to that other number, with the thirty-five million plus the five million.

First, is he confusing the number of base-pair differences with the number of fixed mutations that must have occurred? As I understand this, a single mutation can change many base pairs. It appears that throughout the rest of his argument, when he says "fixed mutations", he is really talking about fixed base-pair changes.

Maximal mutations - Vox Popoli

He averages several estimates the time since the Chimpanzee-Human Last Common Ancestor (CHLCA) to get 9 million years. Using twenty years per generation, he calculates that there have been 450,000 chimp and human generations since the CHLCA. He then uses an average rate of 1600 generations per "mutation" (single base-pair change) to conclude that there could only be at most 562 "total fixed mutations" in this time (the 125 figure appears to be from using 4 million instead of 9 million years). His justification for using the 1600 rate doesn't make much sense to me.

He is citing a Nature article that I don't have access to:
Human mutation rate revealed - Nature

Source: Sequencing of 19 whole genomes detected 25 mutations that were fixed in the 40,000 generations of the experiment.
NATURE, 2009

Biologists REALLY Don't Get Math - Vox Popoli

What JFG is claiming in his clarification is that because sexual reproduction CAN result in faster gene fixation than bacterial cloning when measured in terms of generations, it is possible for the fixation rate to be high enough to account for genetic divergence between chimps and humans since the LCHCA, as well as every other observed genetic divergence over time.

However, this is not true. What JFG and his followers forgot to take into account is that it is not generations that are relevant here, but time. And although bacterial cloning may – or may not – be slower to fixate a gene across a population than sexual reproduction in generational terms, bacteria also have considerably more generations over a given period of time than humans do.

I used the bacterial model in order to demonstrate that the case for natural selection was impossible even when applying the fastest possible fixation rates ever observed.

This is 1600 generations per "mutation". How does the number of generations that bacteria have factor into this number that is supposedly being very generous? He is trying to justify using the rate ("generations per fixed mutation") of bacteria for humans and chimpanzees because he claims it is faster due to bacteria having many more generations for a given period of time, but then uses 20 years per generation to calculate the "maximum fixed mutations". Am I missing something or does this just not make any sense?

Also, is 1600 "generations per fixed mutation" way too slow of an estimate of this rate for humans and chimpanzees since they reproduce sexually and have a much larger genome than bacteria?

 

[Hans Blaster]

Vox Day claims to have mathematically shown that "evolution is impossible". His argument appears to be that simple math shows that there isn't enough time to account for the genetic differences between humans and chimpanzees.

The argument looks badly flawed to me, but I am not a biologist and am likely missing or misunderstanding some things. My question is if I have correctly identified problems or if I am getting things wrong myself.

Your going to base this on the "work" of a non-scientist and alt-right blogger?

This "work" has not been reviewed in any fashion and is just the transcript of a debate. Debates are really bad ways to do science.

 

[AC83]

Your going to base this on the "work" of a non-scientist and alt-right blogger?

This "work" has not been reviewed in any fashion and is just the transcript of a debate. Debates are really bad ways to do science.

I'm fully aware that this is just a debate and not something reviewed and published in a reputable journal. I also know how silly it is to think that this would go unnoticed by professional biologists for so long if there actually were a problem this simple and this glaring. I'm just trying to make sure that I have correctly identified where this argument goes wrong.

 

[Bradskii]

He then uses an average rate of 1600 generations per "mutation" (single base-pair change) to conclude that there could only be at most 562 "total fixed mutations" in this time (the 125 figure appears to be from using 4 million instead of 9 million years). His justification for using the 1600 rate doesn't make much sense to me.

He is citing a Nature article that I don't have access to:
Human mutation rate revealed - Nature

I don't get where he gets that 1600 generations per mutation either. Maybe he's confusing a mammalian generation (say 20 years) with cell generation (which is when mutations can occur). I can't access the Nature article either, but the precis gives all the information he needs:

'Every time human DNA is passed from one generation to the next it accumulates 100–200 new mutations, according to a DNA-sequencing analysis of the Y chromosome.'

So we have 9 milion years ÷ 20 years per generation = 450,000. Multiply by an average estimate of mutations at 150 and we have 6.75 million, which is short of his figure of 30 million. However...we aren't looking at one single line of descent and that 150 mutations. If there are 1000 members of a species in each generation then the number of mutations in the species is 150 x 1000. Or 150,000 per generation.

Obviously most of those will do nothing and will be lost at some point. But we've gone from his one mutation every 1600 generations (which sounds ridiculously small even to a layman - one mutation every 32,000 years?) to 150,000 mutations per generation in a relatively small group.

The problem is not then how do we get to a 2% difference. It's why is there only a 2% difference.

 

[AC83]

I don't get where he gets that 1600 generations per mutation either.

This is apparently a figure from the Nature article for the number of E-Coli generations per fixed mutation, but I can't verify that. What I don't get is how he tries to justify using this number for humans. He makes it sound like this is being very generous by using "fastest possible fixation rates ever observed". However this is just due to bacteria having many more generations than humans for a given period of time and not the generations per fixed mutation rate.

 

[AC83]

I can't access the Nature article either, but the precis gives all the information he needs:

'Every time human DNA is passed from one generation to the next it accumulates 100–200 new mutations, according to a DNA-sequencing analysis of the Y chromosome.'

So we have 9 milion years ÷ 20 years per generation = 450,000. Multiply by an average estimate of mutations at 150 and we have 6.75 million, which is short of his figure of 30 million. However...we aren't looking at one single line of descent and that 150 mutations. If there are 1000 members of a species in each generation then the number of mutations in the species is 150 x 1000. Or 150,000 per generation.

Obviously most of those will do nothing and will be lost at some point. But we've gone from his one mutation every 1600 generations (which sounds ridiculously small even to a layman - one mutation every 32,000 years?) to 150,000 mutations per generation in a relatively small group.

The problem is not then how do we get to a 2% difference. It's why is there only a 2% difference.

Another thing I should add is that the argument appears to be about fixed mutations and not just any mutations the occur in the species. I'm not sure if terms are being used in the standard way, but I don't think this really addresses what he is trying to get at.

 

[Bradskii]

Another thing I should add is that the argument appears to be about fixed mutations and not just any mutations the occur in the species. I'm not sure if terms are being used in the standard way, but I don't think this really addresses what he is trying to get at.

And the guy is an accountant. He admits he's getting his info from wiki. And as I said, his figures of one mutation every 32,000 years means that there hasn't been one since the upper Paleolithic.

 

[AC83]

And the guy is an accountant.

Where did you get that from? Did you mean economist? He likes to compare evolution disfavorably to economics, but he actually isn't an economist either and just has a bachelor's degree in economics.

He admits he's getting his info from wiki. And as I said, his figures of one mutation every 32,000 years means that there hasn't been one since the upper Paleolithic.

I know it still seems like it must be way off, but it was supposed to only be for mutations that become fixed.

 

[sfs]

I don't know what he's doing and I'm not going to waste my time figuring it out. The easiest comparison to make (and the one that leads to the quoted (and not exactly correct) 98% identity between human and chimpanzee DNA) is between a single copy of the human genome and a single copy of the chimpanzee genome, counting only single-base substitutions. These occur at a rate of roughly 50 per generation. If the split in the lineages occurred ~7 million years ago and that ancestral population was fairly large (for which is evidence), the two copies that we're looking at last shared an ancestor ~9 million years ago. At 25 years per generation, that means each lineage has accumulated mutations for 360,000 generations. At 50 mutations per generation, that means ~18 million mutations should have accumulated in each lineage, for 36 million differences between them. Out of a genome with 3 billion base pairs, that amounts to 1.20% of sites that should differ. The observed difference is 1.23%. Any calculation of the predicted difference that isn't in this ballpark is wrong.

 

[AC83]

The reason why the justification given for the human "fixation" rate figure used doesn't make sense to me can probably be explained a little more clearly than in my first post.

Consider the two equations:

X1 = A1*B1

X2 = A2*B2

Let the A variables represent "generations per fixed mutation" and the B variables represent "years per generation". The X variables represent their product, the years per fixed mutation. The X1 equation is for bacteria and the X2 is for humans.

He is reasoning that 1/X1 is high (due to 1/B1 being high), so he claims that he is being very generous in his estimation of X2 by using the A1 value as an estimate for A2 even though A1 isn't why 1/X1 is high. The actual cause of 1/X1 being high, B1, isn't used in the X2 equation at all. He ends up with a low value for the 1/X2 estimate.

This is not "applying the fastest possible fixation rates ever observed". It is applying the "generations per fixed mutation" value that corresponds to what he claims has the fastest fixation rates ever observed even though this value isn't why that fixation rate is a high as it is. This results in a very low estimate for human fixation rate that makes no sense to use.

 

[Bradskii]

I read that three times. Then I gave up.

 

[AC83]

Well, I guess that I didn't make it as clear as I thought.

 

[Shemjaza]

Does he have a justification for fixation rate for genetic changes not being simultaneous for multiple traits in a population?

 

[AC83]

Does he have a justification for fixation rate for genetic changes not being simultaneous for multiple traits in a population?

It doesn't really matter for the argument he is trying to make. It is just the average rate that he is concerned with.

 

[QvQ]

A puzzle to me:
There were no humans during the time of the dinosaurs so humans evolved within the last 66 million years.

Surely there would have been some life form that was "becoming human."

It would depend on how far from "human, the "Becoming Human" was, as to how many mutations were required to be human DNA.

And the argument that there were no humans in dinosaur times is predicated upon the number of mutations subtracted from human DNA to where human was not identifiable as pre or actual human, DNA yes?

 

[Shemjaza]

A puzzle to me:
There were no humans during the time of the dinosaurs so humans evolved within the last 66 million years.

Surely there would have been some life form that was "becoming human."

It would depend on how far from "human, the "Becoming Human" was, as to how many mutations were required to be human DNA.

And the argument that there were no humans in dinosaur times is predicated upon the number of mutations subtracted from human DNA to where human was not identifiable pre or actual human, DNA yes?

No humans in dinosaur times is based on the evidence for the diversity of mammals in that time... and the evidence for the evolution of primates much later.

There are no evidence for primates 60 million years ago, no evidence for apes 45 million years ago and no hominids 10 million years ago.

We then get into the much more detailed evidence for humans emerging as a variant and eventually last man standing of the hominids.

 

[Shemjaza]

It doesn't really matter for the argument he is trying to make. It is just the average rate that he is concerned with.

Then I don't understand how he's able to determine the rate over a long scale... given he appears to want to use historical genetic evidence to disprove the conclusions of historical genetic evidence?

 

[AC83]

Then I don't understand how he's able to determine the rate over a long scale... given he appears to want to use historical genetic evidence to disprove the conclusions of historical genetic evidence?

He just multiplies an estimate of the years per generation by an estimate of the generations per fixed mutation. As I explained, the problem is that his rational for the number he uses for the later estimate doesn't actually make any sense.

 

[Bradskii]

A puzzle to me:
There were no humans during the time of the dinosaurs so humans evolved within the last 66 million years.

Primates started about that time. The genus Homo about 2.8 million years ago.

 

[sfs]

The reason why the justification given for the human "fixation" rate figure used doesn't make sense to me can probably be explained a little more clearly than in my first post.

Consider the two equations:

X1 = A1*B1

X2 = A2*B2

Let the A variables represent "generations per fixed mutation" and the B variables represent "years per generation". The X variables represent their product, the years per fixed mutation. The X1 equation is for bacteria and the X2 is for humans.

He is reasoning that 1/X1 is high (due to 1/B1 being high), so he claims that he is being very generous in his estimation of X2 by using the A1 value as an estimate for A2 even though A1 isn't why 1/X1 is high. The actual cause of 1/X1 being high, B1, isn't used in the X2 equation at all. He ends up with a low value for the 1/X2 estimate.

This is not "applying the fastest possible fixation rates ever observed". It is applying the "generations per fixed mutation" value that corresponds to what he claims has the fastest fixation rates ever observed even though this value isn't why that fixation rate is a high as it is. This results in a very low estimate for human fixation rate that makes no sense to use.

Using only fixed mutations makes no sense at all to me. Given the size of the current human population and our mutation rate, every single-based substitution that's compatible with life is present at least once in the current population, meaning there are virtually no fixed alleles. Even if that weren't the case, the rate of fixation depends on the effective population size and how that has changed over time, which requires modeling and is subject to substantial uncertainties. Which is why the normal practice is to compare single genome against single genome, as I described above.