I think it is. That would preclude the sort of existing variation you’re talking about. It seems that in the old selectionist/neutralist controversy, Behe comes down on the selectionist side, and so would expect little genetic variation within species.
I wrote “not really” because I asked Bill, not you. It indeed is as we understand it, but not as Bill fails to understand it. Keep it simple.
Which we now measure all the time. How many MYH7 alleles have been found in perfectly healthy people to date?
Has Behe ever qualified this as an expectation, or just assumed it as fact?
He didn’t so much “test” a model as assert one – a model that was heavily criticised.
The fact that his one venture into population genetics was heavily criticised would appear to undercut this claim.
What evidence do you have to substantiate this claim?
To the best of my knowledge, Behe’s undergraduate degree was in Chemistry, and his PhD was in Biochemistry – neither of which would require even a basic understanding of population genetics.
Hi John
Mike’s model is about finding a new function not about how much variation a population can tolerate. His model assumes a duplicated gene exists in a population so he has given the process a head start and this will underestimate waiting time to fixation. His model is about two specific types of functions so this will overestimate waiting time. Both these issues are discussed in his paper.
There are tradeoffs in all models as you know but it appears that Mike and David Snoke did a pretty good job as they got a critique and similar model from Micheal Lynch who is a well regarded population geneticist.
As far as MYH7 alleles found in healthy people this idea is not inconsistent with his model as he allows for certain AA substitution not effecting function.
His model also assumes intermediate novel function-creating mutations will destroy the original function and that such mutations are strongly deleterious, and that only one such solution is possible instead of a larger set of possibilities.
This set of assumptions can’t be generalized to diverged duplicate genes, so his model is of no worth when trying to assess diverged duplicate genes like the ones we find in the Venn diagrams you persist in not understanding.
You’ve always and only ever been wrong about how Behe’s paper relates to the Venn diagrams: It does not relate at all.
2 Likes
It also assumes that we’re looking at a particular locus rather than some locus somewhere in the genome, so this will also overestimate waiting time. Combine the two to wait for a specific function in a particular gene rather than some function or other in some gene or other. Big problem.
1 Like
This is true but the model also underestimates waiting time due to the assumption a specific gene duplication already exists in the population.
The model like all models is not perfect but it represents some of the problems protein evolution as a concept faces.
No, it represents a specific imaginary example not known to have ever been required in the entire history of life on Earth. Behe has not before or ever since shown any example of a protein that would have to evolve in the way he imagined. Not even a synthetic one. And regardless, it’s completely irrelevant to the Venn diagrams you obsess over.
Here’s a paper that’s actually relevant to your Venn diagrams (as they explicitly model the divergence of duplicate genes under the constraint that they are required to remain functional during this divergence), but it does not support your imaginations:
1 Like
The key word is “specific”. It’s a Texas sharpshooter model. “Not perfect” covers a lot of ground and includes, at the lower end, “useless”.
3 Likes
Hi Bill,
You asked:
DID he ever test his model? Why haven’t you answered that?
His model also assumes no polymorphisms for either gene, correct? Is that consistent with reality?
No one is ever waiting for anything, except for you to do some math.
His model has no basis in reality.
No, that means that they did a lousy job and the journal’s peer review failed.
Lynch’s model is not similar. Its assumptions are based on reality.
You’re trying very hard to miss the point. It demolishes his assumption. Also, most of these substitutions DO affect function; that’s how they were found.
The point is that function is a very broad plateau, not a peak, and epistasis is really important. Do you know what epistasis is?
2 Likes
Hi John
He did not test his model empirically which is your point. The closest test of his model is the Lenski experiment.
His model shows that polymorphisms show limited to no effect on waiting times.
I am not sure the point you are making. The math is in both Lynch’s and Behe’s models.
The Lenski experiment did not produce a new gene function by duplication and divergence. It did produce a new function by duplication of a promoter alone.
Why?
Lynch’s assumptions are not all based in reality as we know AA substitutions can be deleterious.
It does not at all. Have you read his paper?
You are treating function as generic it is not.
The question is the overall model conservative. Since he is using very simple adaptions, ultra conservative mutation rates, populations with already duplicated genes and species with very rapid reproduction rates, it is hard to argue the results are not conservative based on real biological populations.
Sure, you can argue that anything is conservative by ignoring all the non-conservative bits. In this particular case you are ignoring the fact that it’s a Texas sharpshooter model.
No, my point is that your claim was objectively false. Why did you make it?
I don’t see how. If Behe is doing science, isn’t it Behe’s responsibility to test his own hypotheses/models?
How so? I notice that you didn’t answer my question. Why?
There is no waiting for anything. Pretending that the new function must be waited for is sneaking in the false assumption that it is essential.
We’re talking about whether Behe’s has any basis in reality.
We aren’t talking about the Lenski experiment, but nice try at evasion.
Then that falsifies Behe’s hypothesis, no?
Because Behe’s model has no basis in reality.
I have. Have you understood it?
Anything but that, so you are lying. You also apparently don’t know what epistasis is.
As a geneticist, I don’t have any such problem.
So what set of duplicate genes are we waiting for Behe’s scenario for mutations in, and which can only emerge in the manner Behe models, in that experiment? That’s right, there is no such set of genes. Nothing Behe models is being in any way tested in the Lenski experiment. You’re just making up crap at this point.
Of course, if Behe’s stance is that his modeled scenario is unlikely to happen in any experiment or in reality, then we all agree. But that’s not a problem for evolution. It’s a problem for his model assumptions.
Where does it show that? That’s right, nowhere does it show that. Behe models waiting for mutations to arise. He doesn’t model a scenario in which the mutations of interest are already present in the population (polymorphisms). That’s because he assumes the function-switching mutations are individually strongly deleterious, so they can never arise individually and be part of the background variation we normally see in real biology. His model basically just waits for a lucky double-mutant to arise in one of two specific genes before either of them suffers a null mutation, which is 800 times more likely in his model. If one of the mutations arises in one of the two duplicate genes, the gene is instantly rendered nonfunctional, which puts the other gene under purifying selection against new mutations.
His scenario is not known to apply to any protein in existence.
No, the Cit+ trait still requires prior permissive mutations elsewhere in the genome. The duplication alone does not result in the Cit+ trait in the ancestral background. Other mutations that happened since the founding strain and the Cit+ trait eventually emerged are actually required to potentiate the Cit+ trait.
So, wrong Bill. Wrong. Like you always are.
Still completely irrelevant to the Venn diagrams you do not understand, and even more irrelevant to Behe. The fact that Behe’s imaginary scenario doesn’t happen in the LTEE doesn’t mean Behe’s imaginary scenario is being tested in the LTEE, any more than the fact that a Volcano does not spawn in my kitchen sink is a test of plate tectonics.
So doubly wrong, Bill.
His model has no factor tied to “reproduction rates”. It’s based on generations in populations of constant size, not absolute units of time.
On the other hand, he’s also assuming intermediate mutations are individually strongly deleterious(instead of neutral, or beneficial), that only one gene is a target for a specific function of which only one solution is possible, instead of any novel function to any gene, and that there is no recombination. It’s hard to argue that’s in any way realistic. It’s completely irrelevant that he assumes a duplicate is already present when all the rest of his assumptions are deliberately designed to make the model fail. But it’s done it’s job well: It’s impressed the hell out of people who can’t reason. He’s provided the breadcrumb you needed to satisfy yourself that his model has some sort of generous assumption. With that in hand you feel you can ignore all the designed-to-fail assumptions he’s put in there too.
But that’s all still irrelevant to the Venn diagrams and to common descent.
3 Likes
Seriously. Behe makes a model with extremely restrictive assumptions that imply evolution is unlikely to go through a series of complicated hoops he imagines. We agree it’s unlikely and, given his assumptions, such a series of events would take very long time to happen.
Then there’s some experiment going on, in which the scenario imagined so as to be very unlikely to happen, doesn’t happen. Wow, nobody saw that coming. And this is supposed to be a problem in what way? It doesn’t support anything you believe about protein evolution in general. We are not disputing that given Behe’s assumptions, the results is unlikely to occur.
We are disputing that Behe’s model is a correct model of how novel protein functions evolve. The fact that the scenario Behe models doesn’t occur in the LTEE is expected. He’s designed it in his imagination so that it is unexpected. That’s the whole point. That’s our point. We, too, expect it to not occur. We just don’t think novel protein functions have to evolve in the manner Behe models.
You’re an idiot to a fantastic degree. Report me!
4 Likes
He’s also off-topic, and has been for more than 500 posts. Report him!
1 Like
Part of the problem appears to be that he doesn’t appear to have sufficient understanding of what the topic is about, or what his red herrings are about, to tell whether they are actually related, or whether they are “off topic”. Trying to keep somebody with such a poor understanding on-topic would appear to be a Labor of Hercules.
2 Likes
This thread is definitely reminiscent of Augeas’s byres.
1 Like
But watch him try anyway.
1 Like
The model is of functions we observe such as ligand binding. Behe did not ignore the non-conservative bits he made other parts of the model conservative to balance out this issue.
The bigger issue is the most sensitive part of the model is the number of changes to reach a new function. This is because of the sequence problem. The data we see from genes believed to be the product of gene duplication like WNT can have differences greater than 100 amino acids.