There doesn’t seem to be any data behind this conclusion.
How does one measure specified complexity? Why is it a problem for evolution?
The proteins you are pointing to are the product of 3-4 billion years of evolution. I don’t see how you can use them to determine the probability of the first DNA polymerase evolving out of sequence that existed 3-4 billion years ago.
Where is the evidence for this claim?
In a population of 100 000 individuals, the time for one neutral mutation to come to fixation is 4* 100 000-> 400 000 generations. Only one mutation. Two concerted mutations is already a problem.
google scholar “human de novo genes” or “human orphan genes”
The time to fixation is 4Ne, where Ne is the effective population size which is always much smaller than the actual population size. You would need to fix that in your maths.
Also, one neutral mutation does not wait for another neutral mutation to reach fixation before it begins to increase in frequency. There are many, many neutral mutations that are at different frequencies at any given time.
If there is a beneficial interaction between a neutral mutation and a new mutation it doesn’t require the neutral mutation to be fixed in the population, so fixation isn’t even a requirement to begin with. One of your references earlier talked about neutral mutations hitchhiking with beneficial mutations, so that’s a pathway as well.
The examples I have seen have evolved from orthologous DNA that is shared between species with the addition of a handful of mutations. I think you are confusing unique gene and unique DNA. Those are not necessarily the same thing. In fact, this is really strong evidence for evolution because it demonstrates how new transcripts can evolve. There is also the question of just how many of these transcripts actually have function.
That is true, and will help a very little bit. The maths were already done
The waiting time problem in a model hominin population.
Sanford J, Brewer W, Smith F, Baumgardner J.Theor Biol Med Model. 2015 Sep 17;12:18. doi: 10.1186/s12976-015-0016-z.
Ola Hössjer, Günter Bechly, Ann Gauger,
On the waiting time until coordinated mutations get fixed in regulatory sequences,
Journal of Theoretical Biology,
Volume 524,
2021,
If this is true, you have a point. If this is only a suggestion, then the ‘evidence’ is fiction
Classic Sharpshooter fallacy:
“The program Mendel’s Accountant realistically simulates the mutation/selection process, and was modified so that a starting string of nucleotides could be specified, and a corresponding target string of nucleotides could be specified.”
If it is a true simulation of evolution then it would select for fitness instead of a prespecified sequence.
Same problem here:
" Transcriptional regulation is perhaps the most studied form of gene regulation, where transcription factor proteins recognize and bind to short subsequences of the regulatory region, so that the expression of the nearby gene changes. These subsequences are called binding sites, and typically they have a length of 6–10 nucleotides. The target of the random walk then consists of all regulatory sequences along which at least one of a few pre-specified binding sites appears somewhere along the sequence, possibly allowing for mismatches at one or a few nucleotides between a binding site and a substring of the regulatory sequence (Stone and Wray, 2001, MacArthur and Brockfield, 2004, Yona et al., 2017)."
They prespecify the needed sequence which is the Sharpshooter fallacy. They also focus on a haploid genome and give short shrift to sexual reproduction which can dramatically alter the maths of the alleged “waiting time” problem.
You may want to check out this Uncommon Descent article:
https://uncommondescent.com/intelligent-design/human-and-chimp-dna-they-really-are-about-98-similar/
In it, the author of the article quotes an email he received from Glenn Williamson:
Sharpshooter fallacy, I didn’t know the term before. However, I think that it is a bit fast to conclude for this fallacy on this information alone. Hossjer as professor Mathematics is pretty aware of sequence space and that you need to include a functional space that is more than one specified sequence. He nicely describes this in his article:
Steinar Thorvaldsen, Ola Hössjer,
Using statistical methods to model the fine-tuning of molecular machines and systems,
Journal of Theoretical Biology,
Volume 501,
2020,
Just using a handful out of the many that are possible is still a problem. As it says in the paper:
"The target of the random walk then consists of all regulatory sequences along which at least one of a few pre-specified binding sites . . . "
A few? That hardly seems sufficient. For example, would the changes in the promoter of the lactase gene in humans that leads to lactase persistence count as a hit in their method?
This also leaves the question of how many “coordinated” mutations are possible in a given genome, how many coordinated mutations are needed in a given evolutionary relationship, and how that would interact with sexual recombination.
Again, I’m only seeing a gap on one side here. And it’s not on the side of rationality.
@T_aquaticus
So what specific differences are you talking about, and in what sense do you think those specific differences can not be produced by common descent and evolution?
What you’re trying to do is ignore what we’ve learned about the genetic and molecular mechanisms of embryonic development. You want to stay with the 19th-century understanding of organisms being plastic, so you can say that it doesn’t look like all that much of a difference to go from gastrulation by involution through a blastopore to gastrulation by ingression via a primitive streak (and maybe suggest how it could be achieved through small morphological changes). But that’s an evolution of the gaps! You’re relying on what we don’t yet know to give a loophole for evolution to act. Currently, we may not be able to specify exactly what genomic changes would be required to make changes such as this in development, but it will come.
@T_aquaticus
If we are talking about the human genome, then the vast majority of mutations are neutral. Only about 10% of the human genome shows evidence of selection against deleterious mutations. The rest is evolving at a rate consistent with neutral fixation.
What is your evidence to support that?
@T_aquaticus
That’s not what that paper says. What that paper is saying is that neutral mutations will only evolve neutrally in 5% of the human genome.
Yes it is. In the section you quote it might only be implicit, but later (penultimate para) they say explicitly:
Together with the restriction to regions with high recombination rates, this rule narrows down the fraction of the genome that is neutrally evolving to 5%.
@T_aquaticus
Epistasis. Mutations that are initially neutral can interact with future mutations and result in a beneficial change.
Of course the expression of some genes is affected by neighbouring (or further afield) sequences. It’s part of how development works. But what you’re relying on is that neutral mutations will work in this way and have a constructive effect. In theory it can happen, but only with a probability of something like 1 in 10^18 if two specific bases are involved, in 10^27 if three, etc. You are relying on good luck (or magic).
@T_aquaticus
The only hole is the one you have invented.
I’ve not invented it, but I do recognise it.
@T_aquaticus
It has nothing to do with what I want. We can see the mutations that separate genomes, and those are the ones responsible for the differences in morphology. We know that neutral mutations can reach fixation, and we also know that beneficial phenotypes that result from the interaction of two or more mutations will be selected for. We can observe that they are useful. There’s no Sharpshooter fallacy here.
The key phrase here is ‘we also know that beneficial phenotypes that result from the interaction of two or more mutations will be selected for’. It’s true, but the chance of the right mutations that will give a beneficial interaction is very low, as I’ve just said.
@T_aquaticus
You are still pretending as if there is only one possible beneficial mutation in any given genome.
No, what I was illustrating with the dice is that as the probability decreases you need to roll more dice or more often.
For the example of resistance, there are only a few possible beneficial mutations, all in the same gene so far as I’m aware. For the vertebrate genome there will of course be many locations where beneficial mutations could in theory occur. Eg current estimate is up to 25,000 genes for humans, and quite likely there’ll be several possible locations for each of these. But to get a constructive change to embryonic development, at a minimum you’d need to make a constructive change to an existing control sequence, probably to several and/or the genes of corresponding transcription factors. This would require many coordinated mutations. So, whilst the number of opportunities increases, the overall chance of making a constructive change decreases. And of course population size is much smaller and generation time is much longer.
@T_aquaticus
It will also be susceptible to fixation.
I think you may misunderstand what fixation means in population genetics. It only means that a mutation becomes a normal part of a population’s genome (close to, if not at 100%).
gene fixation
The condition in which a particular allele becomes the only one that is present in a population, because of either natural selection or genetic drift. [Oxford reference]
It does not mean that it is then immune to further change (like Dawkins’ METHINKS…). However, I see from Wikipedia’s entry that maybe others share this misperception.
In population genetics, fixation is the change in a gene pool from a situation where there exists at least two variants of a particular gene (allele) in a given population to a situation where only one of the alleles remains. In the absence of mutation or heterozygote advantage, any allele must eventually be lost completely from the population or fixed (permanently established at 100% frequency in the population). [Fixation (population genetics) - Wikipedia]
‘Permanently established’ is wrong. If a specific location presents an advantage then natural selection will offer some protection. But a neutral location will not benefit from this, and will be as susceptible to change as any other location.
How many “coordinated” mutations are possible? I would say: Three?
How many coordinated mutations are needed? For a chimpanzee-common ancestor to create a human I would say: a million times 50 mutations coordinated? But you may call this a sharpshooter fallacy and you think that potentially millions of different “human” species could have evolved from our common ancestor, but sadly, it did not happen.
Of course, sexual recombination would have influence, but it is not easy to understand what this influence would be.
Dear Taq, there seems to be a fundamental difference in our basal intuitions about how easy mutations lead to new functions. You think that the possibilities are many, and I think that they are very scarce. Already the origin of one protein with completely new function is difficult. In ‘normal’ biology except evolutionary theory, I don’t see scientists that seriously take the appearance of new functions into account. It is mostly loss of information, loss of regulation, damage and so on. Selection is continuously busy to get rid of that noise. We see that in bacteria. If we see the presence of a new type antibiotic molecule or (at the other side) an antibiotic degrading molecule, nobody thinks that it originated recently and we don’t expect that raising in our laboratory, it was already there at one place on earth (soil in China), but we only discovered it recently.
In our lab, some years ago, colleagues wanted to use aptamers for creating tests to detect animal diseases. Publications with regards to aptamers were promising. We used a protocol for generating random oligo nucleotides aimed to bind certain molecules. You produce huge numbers of random oligo nucleotides and then go through an iterative selection process to select the molecules that bind the target. The only function needed was: binding. The results were very disappointing. And after some years, we stopped the project.
So what’s the gap in rationality that only irrationality can fill again?
I’m not ignoring anything. I am asking for specific examples of DNA differences that result in differences in embryonic development, and why you think those DNA differences can not be produced by evolution. I’m actually asking to see the information, not ignore it.
One example:
There are several studies that have looked at sequence conservation and signals of selection within the human genome, and they all seem to come in at ~5-10% of the human genome.
They say explicitly that neutral mutations only evolve neutrally in 5% of the genome. It isn’t saying that neutral mutations can only occur in 5% of the human genome. Neutral mutations and evolving neutrally are two different things.
What the article explicitly states is that neutral mutations will not evolve neutrally due to many different mechanisms. Again, it is NOT saying that only 5% of the genome can have neutral mutations. For example:
It explicitly states that neutral mutations can evolve non-neutrally.
Why would it have to be two specific bases? Do you think there is only one such interaction possible in the human genome? If so, I would love to see that evidence.
For example, how many combinations of two mutations in the modern human genome can result in a beneficial phenotype? Don’t you have to know this number before you can calculate probabilities of one such interaction being found?
The problem is your calculated probability. You are also ignoring the parallel nature of sexual reproduction which can combine beneficial mutations from separate genetic backgrounds into the same genetic background. For example, let’s say there are 100 possible beneficial mutations that are not linked (i.e. are far enough apart in the genome to allow recombination). The odds of getting these 100 mutations is not 100 times the probability of getting 1 of them. Rather, these 100 mutations will occur in different individuals within the population. As these beneficial mutations increase in frequency you will start to get an accumulation of beneficial mutations due to sexual reproduction. This greatly increases rate at which evolution occurs.
Is that true for every single adaptation? I see no reason why it would be.
Examples?
If none of these coordinated mutations for development can even be cited, then what’s the point?
I know what it means. What I was replying to is your insistence that neutral mutations can only be lost. I am pointing out that they can also be fixed.
Neutral mutations can also reach fixation.
There are only three possible combinations of mutations in a given genome that can produce a beneficial phenotype? How did you calculate that?
How did you determine that they are coordinated?
What brand new proteins do humans have that chimps do not?
Why can’t you get a completely new function from a limited number of mutations in an existing functional protein? How do you determine if a function is completely new to begin with?
I’ve done the same thing to test protease specificity using FRET peptides with random amino acid sequences, so I’m familiar with the general use of these types of libraries.
Of course, DNA binding is just one function out of millions of possible functions. There are many functional proteins in humans that don’t bind DNA. You would also have to qualify the level of binding that you were looking at, and compare that to what might be functional in early life or even in modern life.
In other words can anyone point to an actual gap in the four billion years of evolution aka life, nature, that evolution, aka life, nature can’t fill?
(Beyond cosmic rays providentially altering DNA and embryonic development, there doesn’t appear to have been any mention of epigenetic change here.) The Christians arguing here for some kind of capital ‘I’ ID and irreducible complexity and any other ‘gap’ are only arguing from incredulity as far as I can tell.
“…the most common mutations, transitions, are not really ‘copying errors,’ because the keto-enol transition of the base is driving them and the polymerase is working correctly. So if you’d like, that can be seen as providence more than chance.”
– a Christian Molecular Biologist
Not having accepted evolutionary science since youth and YECism, and having endorsed OEC (old-earth creationism) for a long time (decades), that statement was highly influential in the change in my thinking (another was my nephrectomy not a very long time earlier). I’m okay with God being sovereign over abiogenesis as well, with no exceptions to the natural order required – all laws intact, just exceptional and extraordinary timing and placing. Kinda like the same person winning five independent lotteries and being the only one to have bought any tickets, and only one ticket in each, right? (I won’t tag those who know the rhetorical question is directed to them, although I started to for one of you. ; - )
how can you quote the biologos article on bacterial flagellum when Michael Behee (as one of an ever increasing number of creation scientists) exposed this problem in the evolutionary theory and have shown that its a very problematic example for evolutionists?
No it’s not. It’s entirely his problem and he utterly fails to make it evolution’s, life’s, nature’s problem. Apart from nature having come up with Michael Behe. And you. And Layton and Erik the Vet. Which is normal in evolution.
Because my Bayesian priors led me to put up blinders to comprehend anything contradictory to the oath I took to worship at the altar of evolution.
“Let your conversation be always full of grace, seasoned with salt, so that you may know how to answer everyone.” -Colossians 4:6
This is a place for gracious dialogue about science and faith. Please read our FAQ/Guidelines before posting.