Behe’s model is not about bacteria - it’s about all organisms.
That post I keep referring you to - the one where phage lambda generates a new binding site with four mutations - is really quite informative on this issue. I think it would greatly help you understand what Behe is arguing, and why his argument just does not work.
And do you realize that Cornelius raised his objections here, and that I can raise my objections here as well?
And do you have an opinion on the massive contradiction between pointing out that utterly specific, high-affinity antibody production, which recognizes virtually every antigen in existence, explores only a tiny amount of sequence space, and claiming that evolution must explore vast amounts of sequence space to produce a much lower-affinity binding site?
[quote=“Billcole, post:156, topic:9418”]
Are you saying the evolution only happens one mutation at a time?[/quote]
Behe claims that both must be simultaneous. That’s wrong.
Has he demonstrated a requirement, or merely asserted it?
[quote]One mutation at a time only works if the mutations find the right positions.
[/quote]Mutations “finding” positions is not a useful metaphor.
Hi Bill - short answer, no - there is no evidence to support that idea. Viruses have a very high mutation rate as it is, and as you will see from the paper, they observe lots of other mutations in their population of viruses.
Somatic hypermutation, again, does not specify what amino acid changes might be made - it only biases mutations towards a particular part of the B cell receptor coding sequence.
"Unfortunately this example also fails to demonstrate Venema’s claim of “evolution producing a new protein-protein binding event.”
Are you saying this (a) paper does not demonstrate a new protein-protein binding event, or (b) it was not evolution that did it?
Your critique seems to be that this paper is a freak, one-off event. Is that what you are saying? These sorts of things can’t happen, but even though we observed it happen, it still isn’t meaningful?
The full column is not required to understand the empirical substance of Cornelius’s massive contradiction, Eddie.
Here are the empirical points:
We know how many variants of antibodies begin the process when an organism is challenged with new, unknown antigens. We know how tiny the proportion of sequence space occupied by them actually is.
We know that antibody recognition occurs with far higher affinity and specificity than most biochemical interactions.
We know that we can drastically reduce the number of initial variants experimentally, and we know that we have to perform an insane reduction to put a dent in the ability to recognize new antigens.
Thus, antibody production by variation and selection is far more highly constrained than most of biology, yet it works beautifully in real time.
I think he’s just trying to throw anything he can at it, since being corrected on his false claim that “Specific binding requires specific sequences” made no change in his conclusion.
Well the paper nicely illustrates the “all or none” nature of protein-protein binding sequences, in the sense that those viral J proteins with anything less than the four types of mutations were unable to sufficiently bind to OmpF, but those with all four did. No big surprise there, but a good illustration.
What they also found is that, nonetheless, the viral J proteins did sometimes have those 4 mutation types, but that was only because subsets of those 4 did, fortuitously, increase the virus fitness for a different function. As the paper explains, that was “lucky”. It was also dependent on the host having certain mutations, as well. So it simply is not accurate to say this solves the general problem of evolving protein-protein binding sequences. It would solve that problem only if we could show that there is some good reason to believe such fortuitous, alternate functions, are the rule. I.E., That there will always be some other function in the pathway to the protein-protein binding sequence. The study provides no support for such an conclusion, and even admits it was a fortuitous circumstance. That is not to say it is impossible, but it is not a finding of the study.
Now it also is important to understand that the entire experiment depended on a very artificial, clever, design that set the problem up. This was not even an attempt to simulate any kind of realistic evolutionary environment. There is no reason to think that would be representative realistic scenarios.
This is not at all to say the research isn’t good. Clearly it was a fine piece of work. But your inference of it demonstrating “evolution producing a new protein-protein binding event” is not warranted.
Yet the protein-protein binding by antibodies is anything but “all or none,” as they are produced by recursive rounds of selection, starting with a library that occupies an incredibly tiny proportion of sequence space. Despite that, your body will respond by producing antibodies that bind avidly and specifically to any unknown antigens you encounter.
That shows that highly specific binding is easy to find in sequence space.
Your claim that “The problem with evolving protein-protein binding is that too much gene sequence complexity is required to achieve the needed binding affinity,” is empirically, objectively false. Very little complexity is required.
Hi Ben
Your analogy appears to me to compare Michael Jordan (in his prime ) demonstrating he can jump up and touch the top of the empire state building because he has recorded a 47 inch vertical jump. The jump is impressive but it won’t support the extraordinary claim.
This is a not a meaningful statement until you define the size of the sequence space. Since the space grows exponentially with every added component.
To be clear, then - are you saying this was not evolution that produced this new binding site? How is it not evolution? It happened, we observed it, it was selected for.
You’ll notice that the authors of the study say it was fortuitous to observe this on the timescale of this experiment. On longer timescales, are you suggesting that neutral mutations will never combine to achieve new binding specificities? Why not?
My discussing that paper was in response to Behe claiming that more than two or three mutations to form a new binding site was beyond the reach of evolution to achieve. Can you see why this example shows that Behe’s claim is unfounded?
) demonstrating he can jump up and touch the top of the empire state building because he has recorded a 47 inch vertical jump. The jump is impressive but it won’t support the extraordinary claim.