Methinks it is sort-of like two weasels

There’s zero evidence that evolving a new binding site requires deleterious mutations. You can’t take the fact that most mutations (as a proportion of all mutations) are deleterious to imply they are required to generate a new binding site. That just doesn’t follow. Deleterious mutations will be selected against, almost always leaving a pathway through the neutral and beneficial mutations to new functions.

I already cited you a paper that shows direct experimental evolution of a novel protein-protein interaction, and at no point in this experiment did this require that deleterious mutations accumulate in the proteins:

They literally reconstructed the ancestors of some extant proteins, and tried to re-evolve the functions these proteins now have in extant life. They found they were able to do this over and over, and that the novel protein-protein binding function evolved by a different pathway literally every time.

They even went so far as to make changes to the ancestral protein and tried to re-evolve the new function from a different location in sequence space, and found that they were still able to consistently re-evolve the new function, but that this simply made the descendant protein even more different in sequence from the historical outcome. Empirically proving, by experiment, that there are innumerable pathways open by which protein-protein binding sites can evolve, and that the historical outcome we happen to have is just a contingent fact of history.

This idea you have in your head, which you’ve been told by creationist propagandists, simply does not correspond to observational reality. New protein functions can evolve quickly and easily under selection, through beneficial and neutral mutations, and there are hundreds, to thousands, some times millions of different pathways through which proteins can both climb and neutrally wander into those new functions.

7 Likes

If every step is selectable! But insert non-selectable steps in the mix, and evolution bogs down.

No, but it does help if you’re on a nearly 2000-step journey. And if I’m not mistaken, I think they require each step to be fixed in the population.

“For example, if the chances that a cell will be resistant to a given drug are one in a million, and if there are a hundred million parasitic cells in a patient, then almost certainly there will be some resistant cells in the patient.” (The Edge of Evolution, p. 55)

Taking into account existing variation, here.

That’s what Summers et al. found out.

I don’t recall saying that…

But as shown, Behe doesn’t ignore existing variation. And without pressure from chloroquine, the variants with resistance disappear, so it appears that resistance does have to appear de novo.

And why are you saying that existing variation, recombination, and sexual reproduction make searching through shape space fast and efficient?

And that’s fine, only it’s not part of the process of evolution of DNA.

Behe doesn’t make a direct correspondence with protein-protein binding and PfCRT evolution, though. Have you read his argument?

Yes it is. It’s part of the process of evolution of the DNA encoding the antibodies expressed by those immune cells.

2 Likes

You are mistaken.

4 Likes

Neutral mutations can reach fixation without needing selection.

The human genome is chocked full of mutations that weren’t selected for.

2 Likes

Mentioning it, but not including it in any of his simplistic calculations. Interestingly, you completely contradicted this claim in the other thread:

No, you are misrepresenting the paper. Did they assay all the mutants known in 2014? Have other mutants been identified since then?

Convenient. There’s a search function at the top of the page.

Your claim is spectacularly, totally, objectively false. Why do you make such claims?

Mentioning it once doesn’t constitute including it in his simplistic calculations.

Isn’t that precisely what happens in the phases of the Plasmodium life cycle in mosquitoes and rats? Is the fitness of these mutants higher or lower than wild-type in the absence of chloroquine, Lee?

Please show me where anything resembling that is included in Behe’s calculation.

Because they create variation without new mutations.

Why do Summers et al. refer to haplotypes, but Behe never uses the word in the book? That should tell you something.

Why does Behe never mention epistasis?

1 Like

I was responding to your claim that he does, remember?

That’s ludicrous.

1 Like

He made the false claim in two places. Here’s the other one:

No new gizmos or basic machinery. There have been no reports of new viral protein-protein interactions developing in an infected cell due to mutations in HIV proteins. [21]

Reference 21 is
Wang, J
Trends Biochem Sci
Volume 27, Issue 3, 1 March 2002, Pages 122-126
Protein recognition by cell surface receptors: physiological receptors versus virus interactions

It has nothing to do with Behe’s claim.

1 Like

But again, evolution had a chance to use existing variation in evolving chloroquine resistance, Behe thus did not exclude it in his 1 in 10^20 number,

Meaning already-existing protein-protein interactions, sorry that wasn’t clear.

“We measured the ability of more than 100 variants of PfCRT to transport CQ…” So they got a good sample, even though it may not be exhaustive.

The fitness of the mutants is lower than wild-type in the absence of chloroquine, would be the claim. So then treating the two mutations in each known path as deleterious singly, would explain why Behe states that both mutations need to occur together, to get the beneficial result.

I don’t know why this is important–but Behe references a paper that mentions haplotypes, in reference to sickle-cell mutations.

Because evolution had this at its disposal, during the time chloroquine was in widespread use.

But I meant Behe’s argument, using the arrival of chloroquine resistance as what he calls a CCC (chloroquine-complexity cluster), and comparing that probability to the probability of getting a new protein-protein binding site.

I would assume HIV is not referenced in this paper about viruses and protein-protein interactions. That would then merit mention as a reference.

That’s fine, I mean non-selectable steps in a path to a new structure.

How so, though?

“The response obtained in each generation would then be R = 0.00005m, which means that the small variation and weak selection cause a change of only 0.005% per generation.”

Sounds like they mean the population changes .005% per generation.

Sure, Behe is just saying they are probably deleterious.

So the ancestral proteins had the same function as the extant ones? If so, then the mutations were simply improvements, and the binding site was already there.

So evolutionary steps are more than 99% probable? I think I’m being generous.

But I think the probability of 1980 successes is just .99^1980.

Those steps already exist in the human genome and in the genome of every species.

2 Likes

They do, but what’s changing is the population mean for some physical characteristics. The underlying genetics, as I’ve tried to explain, does not require fixation of alleles, much less fixation of one before the next can begin.

5 Likes

He did exclude it, because Behe falsely claims that everything has to come from new mutations. That necessarily excludes existing variation.

Most new interactions begin as already-existing ones.

How good? How exhaustive? What are the numbers? You don’t have the slightest idea, but you’ll say anything to defend it. Why?

No, the fitness could be the same. You’re missing the point by a mile.

Or, intermittent selection explains the low frequency of fixation.

So what? What does the term mean? What if two different mutations are on different alleles? Let’s see, might there a way to get them together in the same allele? You won’t learn this from Behe.

But he never considers it as a mechanism. Why? Does he think that his readers aren’t very smart?

Doing that is academic misconduct. Look at how reflexively you make up silly reasons–even when you already know that Behe’s claim of zero is objectively false!

2 Likes

No, OVERLAPPING functions. Protein function, especially binding, is not digital. Partially-overlapping function is not a characteristic of designed systems, but it is predicted by evolutionary mechanisms, particularly the ones that Behe pretends do not exist.

3 Likes