I would note that many people disagree with that particular point and would claim that much of the phenotypic variation in a population is neutral too, though they would probably agree that a mutation that affects phenotype is less likely to be neutral than one that doesn’t.
1 Like
How did humans get the intellect to adapt to so many different environments? Are you saying that capacity is not in the genetic code? How many beneficial mutations does it take from taking a stick to get ants and termites out of their nests to making computers and spacecraft? Then show us how a lineage of humans accumulated these beneficial mutations. Be sure to show your math.
It’s those beneficial mutations that need to be accounted for. And you don’t have very many replications available to do your math, only about a billion up to 10k years ago.
And if population size is not the measure of reproductive fitness, then what is? Humans can survive and reproduce in almost every environment on earth. Humans adapt using their intellect and if that capability is not in the DNA, where is it? Chimpanzees don’t have that reproductive fitness.
Humans have adapted to almost every niche on earth. And this ability to adapt is based on intellect. Do you think that intellect is an expression of genotype? And if humans were to fight with chimpanzees, they wouldn’t do it hand to hand (at least not willingly). If humans took the habitat occupied by chimpanzees, human intellect could achieve far higher populations than the 300,000 that chimpanzees have achieved. Chimpanzees don’t have the intellect to understand farming.
I started this thread because Joshua made a video saying that the genetic differences between humans and chimpanzees can be explained by neutral evolution. I don’t agree with that claim either. If one makes the claim that humans and chimpanzees arose from a common ancestor, then the correct genetic transformational accounting rules need to be provided to explain those differences in anatomy and physiology and neutral evolution doesn’t give that explanation.
When I listened to your 2018 Fisher lecture, you made a brief comment about entropy. A simple definition of entropy is it the measure of disorder. In the introduction of my undergraduate thermodynamics text, the author made the following comment; “Entropy like beauty is in the eye of the beholder.” The way I took this statement, it’s like looking at hieroglyphics, without the rosetta stone, the order can’t be seen. DNA evolution in my view is one of the easiest examples of the 2nd law thermodynamics to see. Consider this analogy, I write a computer algorithm that takes a string of characters, runs that string through a loop and each time randomly change a character. For example, let that input string be “Joe_Felsenstein”. The sequence corresponds to something real, there is order (information) in that string. But with every run through the loop, information is degraded, it is becoming less clear what that string corresponds to, and without selection, there will be a random sequence of letters that has no correspondence.
[/quote]
[Mod edit to correct mangled quotes, long after the fact - DE]
What is the relative difference in fitness required to account for the observed difference in population size in ~300k generations?
1 Like
With mutation like that, and no natural selection in the model, yes, that’s what will happen. Should not surprise anyone. So?
4 Likes
But that is exactly what happens and has happened.
We have almost as many pseudogenes as functional genes.
Also, funny how the pattern of pseudogenes in different species matches that predicted by common descent
2 Likes
Human/Pan physiology is really not my thing, but I can reasonably speculate that most of the necessary qualities were already present in the last common ancestor. the difference then are largely quantity - more intellect and greater language capacity, leading to better tools use and passing on of learned behaviors. Most of these would be regulatory changes, I think? Evo-Devo isn’t my thing either.
This is pretty far afield from your OP statement, which was to discuss the flaw … “where Dr. Swamidass tries to explain the genetic differences between humans and chimpanzees using the concept of Distance = Rate x Time, neutral evolution.” I understand that you disagree, but you haven’t justified a flaw. Neutral evolution isn’t supposed to explain differences that allow positive selection (by definition).
But humans have not adapted to every environment on Earth, and remain clawless naked apes. What we have gained is the ability to adapt our environment. This is clearly a big reproductive advantage, but the genetic change change is more subtle. Humans clearly crossed a threshold into a new mode of living where the old rules of fitness no longer apply. To my knowledge no one has pinned down exactly the DNA that enables this, but if I had to guess it is a synergy of small effects, not a massive change in DNA, that makes us human.
3 Likes
If I understand @swamidass correctly, his point was that the simple number of mutations differentiating humans from chimps can be accounted for purely thru neutral mutations, taking into account mutation rate, fixation rate and the time since the MRCA. (He will correct me if I am wrong, I am sure.)
This does not mean that the specific differences between the two lines are entirely due to neutral mutations.
The existence of beneficial mutations, however, only make the creationist case more difficult, since they will be fixed more rapidly and therefore make it even easier to account for the number of mutations occurring over the given time scale.
Does that help?
2 Likes
That’s about right but it’s more than merely the numbers. It’s also relative patterns, such as the fact that X chromosomes are less different than Y chromosomes, etc.
4 Likes
That’s actually an observation, not an assumption. That number comes from direct sequencing of the genomes of parents and their offspring.
I thought I included that in my calculations. You are correct, 180 million replications to get all possible mutations.
One of the earlier attempts at estimating the human mutation rate looked at de novo mutations that gave rise to dominant Mendelian diseases:
How many possible combinations of 2 mutations would produce this outcome? That needs to be part of your calculations. Also, if each of these mutations is beneficial on their own then they can be combined into a single genome through recombination.
Let’s do another back of the envelope calculation. Let’s use the possibly low estimate of 50 mutations per individual per generation, a constant population of just 100,000 humans, and 25 years per generation. That’s 5 million mutations, and over 5 million years we have 200,000 generations. That’s 1 trillion mutations over that time period. We are only separated from chimps by about 35 million substitution mutations, about half of which would have happened in the human lineage. This means we only needed to keep a tiny, tiny fraction of the mutations that did occur in order to produce the genetic differences we see.
5 Likes
Are you saying that the difference in cognition between humans and chimps is not due to the differences in our genomes?
It seems rather obvious that the beneficial mutations are among those genetic differences. We also see a natural process that produces those mutations.
Which genetic differences between humans and chimps could not be produced by the known and observed natural processes that produce mutations?
I would agree with this statement with the caveat “the vast majority of differences between the genomes of humans and chimps is due to neutral drift”. This is true.
4 Likes
You will see several people pointing out the issues with the Texas Sharpshooter fallacy. You seem to be under the impression that the genetic outcome of evolution in any given lineage is the only outcome that was possible, and then you sit back at amazement that evolution could have hit such a tiny target. This is wrong, of course. Evolution could have gone in many directions in the human lineage, and there is every reason to believe that there were many possible genetic changes that could have produced similar outcomes.
A recent example from here at PS is lactase persistence. Mutations allow for lactase production into adulthood instead of shutting down after weening. It turns out that mutations in upstream of the transcription start site of the lactase gene changes its expression in adulthood. There are multiple mutations that do this, as discussed here:
There is absolutely no reason to think that there are just 2 mutations that had to occur together in order for human intelligence to evolve. There is every reason to believe that there are perhaps millions of possible pathways if we include gene duplication events.
2 Likes
A useful way of thinking about the question! Starting with the last common ancestor of Pan and Human, which must have already been a fairly intelligent creature; what is the probability that none of it’s evolutionary descendents would develop a higher order of intelligence? Subtract this probability from 1.0.
2 Likes
We can use Kimura’s absolute fitness equation to approximate that value:
n(t+1)=Wn(t) where n is the population size of a particular genotype at generation t. If we assume W is a constant over generations (which obviously it isn’t but we can get a sense of that particular genotype’s average growth in population size over many generations), we can obtain an equation that estimates W based on initial population size and some population size “G” generations later. Start by computing the population size at generation 2.
n(2)=Wn(1) then compute the population size for generation 3
n(3)=Wn(2)=WWn(1)=W^2n(1) then compute the population size for generation 4
n(4)=Wn(3)=WW^2n(1)=W^3n(1) we can continue this process for G generation and obtain the following equation:
n(G)=W^(G-1)*n(1) or
W^(G-1)=n(G)/n(1) raise both sides to the 1/(G-1) power gives:
W = (n(G)/n(1))^(1/(G-1))
Assume that humans and chimpanzees at generation 1 have a population of 2 and at generation 300k, 7E9, and 300,000 respectively for each.
1.0000397289
1.0000732564
But consider that most of the human population growth has occurred in the last 10k years or about 1k generations. Then set n(1)=5,000,000, n(1000) = 7e9 and the absolute fitness for humans during that interval gives a W value 1.0072778346.
If you think I’m not being fair by considering world-wide populations, just consider Tanzania and Senegal, ~75,000,000 humans vs 300,000 chimps
That’s the analogous computation that the Jukes-Cantor and your F81 model are doing. You are computing the rate of disorder of a particular site in a genome. You start the Jukes-Cantor model with one of the state frequencies as 1 and the three other states as 0. The random walk starts and the state that initially was frequency 1 slowly decreases (rate dependent on the mutation rate) and the other 3 states initially at 0 frequency slowly increase. Because of the symmetry of the Jukes-Cantor model all the states converge on the value of 0.25. That is the eigenvector for this particular Markov matrix when the frequencies of the possible states stop changing. That can only happen if that particular site in the genome is not under selection. But this is not how real populations do their random walks. This is why I encourage you to study experimental models such as the Kishony and Lenski experiments, especially the Kishony experiment.
As Kishony’s populations grow, there are many variants taking their own particular random walks. The question then becomes, how large must the population grow for there to be at least one variant getting a particular mutation at the correct site to give improved fitness if the frequency of that particular state initially is 0. It requires taking population size into account in the Jukes-Cantor (or your model as well) to obtain that theoretical value. I encourage you to read the paper that explains this in detail:
The Kishony Mega-Plate Experiment, a Markov Process
The environment determines which one of these random walks give improved fitness.
Do the math but do it correctly.
Learn what a lineage must do to accumulate a set of beneficial mutations. Using T_aquaticus’s mutation rate, it takes 90,000,000 replications for that first beneficial mutation. Then there is the matter of the second beneficial mutation occurring on some member with the first beneficial mutation.
[quote=Kleinman]
You are assuming a mutation rate of 1.67e-8 which some might say is a little high but ok, let’s run with your number.
Are the measurements done with germ cells or somatic cells? Considering that the human body contains 20-30 trillion cells, that’s a lot of mutations that can be accumulated in a lifetime. Perhaps the measurement being done take this into account but I think you will find that the mutation rate is not going to be the problem, it’s going to be the multiplication rule.
[quote=Kleinman]
60 million replications of the genome will give a mutation at every possible site in the genome but you have to multiply that number by 3 to get every possible base substitution which gives 1.8e8 replications of that genome.
That’s ok, you are on the right track, it can be hard to see arithmetic errors.
Start with the simplest case of only a single possible evolutionary trajectory. Once you have that figured out, if you can’t figure out how to do the math for multiple possible evolutionary trajectories, I’ll show you how to do that math. To get a sense of this, consider the Kishony experiment (I know, no recombination). Each lineage that adapts to the drug selection pressure still has to create its own colonies of a billion bacteria for each evolutionary step. With the mutation rate you want to use, that’s only 90,000,000 replications for each lineage at each evolutionary step of improved fitness.
And again, my suggestion is to leave recombination out and once you understand the simpler case, we can consider under which recombination can contribute to improved fitness.
What I’m saying is that Chimps don’t have the cognitive power that humans have and that humans have greater reproductive fitness than chimps. And I’m proposing a thought experiment that it only takes 2 mutations to have the intellectual capability to do farming. You have shown that with 90,000,000 replications, you will have 1 member of the population with one of the needed mutations, call it “A” and another member of the population that already has the second mutation, call it “B”. For now, neglect recombination. How many replications of the A variant are needed to get a descendant with the B mutation or how many replications are needed for a B variant to get an A mutation?
That’s the point, I’m trying to get you to see how these natural processes work.
You have actually measured which genetic differences between humans and chimpanzees are beneficial, neutral, and detrimental? Forget the thought experiment, tell us which mutations humans have that give their reproductive fitness advantage over chimps and let’s do the math for those mutations.
You have yet to figure what a lineage has to do to travel a single possible evolutionary trajectory. When you do that, the math for multiple possible evolutionary trajectories doesn’t change the probabilities significantly. Again, study the Kishony experiment where he has multiple different lineages taking their own particular evolutionary trajectory to high-level drug resistance. Watch this video and pay attention to the phylogenetic trees drawn in at the end of the video:
The Evolution of Bacteria on a “Mega-Plate” Petri Dish (Kishony Lab)
Each node in those phylogenetic trees represents a colony of about a billion bacteria. Every lineage must address their own personal instance of the multiplication rule.
NOTE for @kleinman: the quote tags need to be on a line all by themselves.
[QUOTE=so and so]
text here
[/QUOTE]Yes, it does change the probabilities. Drastically.
You are committing the Texas Sharpshooter fallacy. You are pretending as if humans and their exact genome were the target at the beginning. They weren’t the target.
Also, you are claiming that there is only one single genetic pathway for producing intelligence, and that it can only precede one mutation at a time. This is false. There are multiple changes to multiple genes that could each be beneficial on their own. These can be combined through sexual reproduction.
What you keep ignoring is sexual reproduction, something that isn’t present in the Kishony experiment. Let’s say bacteria are sexually reproducing. If the bacteria carrying resistance for one antibiotic are able to mate with bacteria carrying resistance to a different antibiotic you would get offspring that could carry both antibiotic resistance genes. Do you understand why this changes the probabilities?
These are done with somatic cells in different tissues. They are able to differentiate between somatic mutations and the mutations the individuals were born with.
That’s not how the evolution of intelligence works.
If there are multiple possible pathways then the odds of a feature evolving are much higher.
The problem is that your entire argument falls apart when sexual reproduction is considered.
So says the person ignoring sexual recombination.
Do you agree or disagree that among the mutations that separate chimps and humans are mutations that have been beneficial in the human lineage? It seems rather obvious that some of those mutations are beneficial and are responsible for the cognitive differences between us and chimps.
3 Likes
I was trying to fix those, but there seems to be something else wrong. I suspect hidden characters.
edit: closer …
edit2: Got it! It was like playing Whack-a-mole with broken tags!
1 Like
This is very puzzling. We were discussing humans and chimps. Lots of people were taking exception to Kleinman’s statements about their relative population sizes being relevant to measuring their relative fitnesses. I was responding to people about whether the large number of sites undergoing neutral change implied that the ones that affect phenotypes are changing neutrally (short answer: no). Kleinman keeps trying to drag in entropy and push the discussion towards Kishony’s stuff.
Sorry, Charlie.
9 Likes
Which sounds fairly easy to account for…
Which can be explained by a reasonably small change that specifically results in more direct competition, i.e. habitat destruction and bushmeat.
Obviously none of this actually matters, because you still refuse to understand how population dynamics actually work, but it is nice to see you’re still wrong any way.
4 Likes
I thought it was telling that he postulated an original human population of two. Now where could that number have come from?
7 Likes
(Does the maths)
I doubt anyone would agree that 1.00728 is “vastly greater” than 1 in any context.
3 Likes
I can’t recall if this has already been addressed: One assumption you seem to be working under is that there would have to have been more beneficial mutations fixed in the human genome than in the chimp genome in the time since our divergence from our MRCA.
If so, that is another thing you have gotten wrong.
2 Likes