Chance and Providence (reprised)

Conversation Side Conversation
, ,
103

Okay, but why “critical”? Do you mean to say no other amino acid could substitute for it at all without being lethal(instead of just having considerably lower relative fitness), and if so how do you know that?

When you observe the DNA it is mutating at a much higher rate than the AA’s which indicates that purifying selection is eliminating organisms with that specific AA mutation.

Yes that is the evolutionary explanation for conservation in protein sequence. Synonymous mutations are allowed because selection operates primarily at the level of amino acid sequence, but not that much at the level of codons.

Now the question is, are the amino acids we see as unchanging between many different species and which we therefore infer are “conserved” over long timescales, conserved because other amino acids in that place would be lethal to the organism, or do they merely have lower relative fitness? How do we tell the difference?

104

He doesn’t actually have to do that. YOU are the one making the claim that conservation implies “limited substitutability”.

As explained, we’d like to know how you tell the difference between amino acids being selected against because they have lower relative fitness, and amino acids being selected against because they’re lethal(and thus cannot be substituted).

There is really no hill in this application.

How do you know? You’re making this claim now, so how do you know? How do you know the extant sequence didn’t evolve by selection metaphorically climbing up some hill where amino acids got added, or replaced and fixed because they were more beneficial?

We have at least two competing explanations here, the hill-climbing type explanation where the extant sequence evolved by other amino acids getting replaced because they had lower relative fitness, and the one you are proposing where the other amino acids are selected against because they simply can’t be substituted at all and so, presumably, you think this means the sequence must have been originally designed.

So, how do you know? You’ve claimed there really is no hill. How do you know?

105

If you read my response I clearly stated that it is possible for substitutions where we see preservation but also there are possible limits where we don’t see preservation. There is a very large gap between the resulting information in bits and a number that might be achievable by known mechanisms.

As I stated before the application here requires precision and involves about 200 proteins. How would you picture a step by step process performing this function given a genome full of introns. We need to splice out introns at a rate that allows protein production and that allows an animal to be built and sustained.

I don’t think that there are fitness gradients when you talk about a function that is so mission critical. It works or it does not. This is a strong candidate for most the preservation being caused by purifying selection.

It needs to accurately splice out introns or you cannot make proteins. There is no almost application here such as you might have in a stand alone enzyme with weak catalytic activity.

Describe the hill it is climbing keeping in mind this is one out of 200 proteins in the complex. Describe how the sequence found the hill. Describe how the hill it found had a high level of preservation at the top. Describe how the other proteins in gpuccio’s chart were able to find optimized hills. Describe why preservation in the chart was scaling with sequence length.

I don’t expect you to answer all these just food for thought.

106

Please cite whatever you’re talking about.

Once again, that’s your inability to understand that there is indeed a hill. Just because selection maintains species at the top of the hill, that doesn’t mean that they didn’t arrive at the top from somewhere else. This is your failure of imagination and understanding.

107

I ran blast this morning of mouse, human, rat and slime mold. 77% alignment. When I ran human, rat and mouse 99.9% alignment. When I ran human against slime mole 77% alignment. For this protein the methods appear to be getting similar results.

You have not explained how a hill applies to removing introns from RNA.
You have not explained how 1 protein of a 200 proteins complex climbs a hill on its own.
For animals separated by over 50 million years there is almost perfect alignment. What are your thoughts here?

108

I don’t think you know how BLAST works. At any rate, why three mammals? How is that a test of anything?

That should be obvious enough. Start with a complex that only removes introns correctly some of the time, causing a certain percentage of garbage translation that wastes energy. As long as there’s enough correct splicing to support the cell, no big deal. Oh, wait. That’s what still happens today.

Well of course it doesn’t, on its own. What?

The protein is highly conserved. It’s at the top of the hill, and going down the hill is selected against. This says nothing about the existence of the hill, as people have tried to explain to you for years now.

109

How do you explain it found this theoretical “hill” in almost infinite search space?

I just ran blast on brr2 for humans mice and rats and it is 99.5% similar so it also found the theoretical peak in almost infinite search space. The only difference here is that rats have 3 additional AAs.

How did a search find these exclusive hills?

How did you get to this point?

110

SF3B1 for humans and mice again 99.9% similar. It found the peak again. With all those theoretical hills to climb it keeps finding the peak in almost infinite search space.

111

That did not answer the question I posed and to which you posted that in response.

Possible limits? Yes, sure, there are possible limits. In fact I’m sure there ARE limits. What we want to know is how you determined what those limits are among multiple options.

I must repeat the question it seems: Do you mean to say no other amino acid could substitute for it at all without being lethal(instead of just having considerably lower relative fitness), and if so how do you know that?

There is a very large gap between the resulting information in bits and a number that might be achievable by known mechanisms.

I have no idea what that means or how it constitutes an answer to my question.

That’s also not an answer to the question I posed to you. Let me quote it again: Now the question is, are the amino acids we see as unchanging between many different species and which we therefore infer are “conserved” over long timescales, conserved because other amino acids in that place would be lethal to the organism, or do they merely have lower relative fitness? How do we tell the difference?

So, again, how do we tell the difference between those two options?

We need to splice out introns at a rate that allows protein production and that allows an animal to be built and sustained.

Sure. So how do we tell the difference between selected against because they had lower fitness, or because they’re lethal?

We already know you believe this. This is your conclusion just stated in different words. What I want to know is how you determine that is actually the case. How do you know that it is so “mission critical” lower fitness isn’t a possible option.

It works or it does not.

That would be true even if there was lower fitness variants. Those lower fitness variants would either work or they wouldn’t. No matter how fast I can ride my bike, it either moves or it doesn’t.

This is a strong candidate for most the preservation being caused by purifying selection.

I agree. But purifying selection against lethal, or just lower fitness substitutions? How do we tell the difference merely from conservation?

Yes, but that says nothing about whether it could be slightly worse at it and still support life.

There is no almost application here such as you might have in a stand alone enzyme with weak catalytic activity.

How do you know that?

Well by definition hills have slopes, that’s the crucial point. With respect to explaining conservation in terms of movement on a fitness landscape, moving downhill means having lower fitness, so here the steepness of the slope is analogous to the strength of selection. So we see relatively little diversity of sequences because that would imply movement down from the top of the hill.

You’re saying there’s no hill, it’s a sort of tiny spike with pretty much vertical walls. How do you know that is the case?

Describe how the sequence found the hill.

I don’t know and I don’t need to know to point out that your haven’t supported your claim that the mere size of sequence space tells us no hill exists.

Describe how the hill it found had a high level of preservation at the top.

It seems to me you’ve explained that yourself: Other sequences are selected against. Now the question is why they are selected against so strongly. Is it because they’re significantly lower fitness, or because they’re lethal?

Describe how the other proteins in gpuccio’s chart were able to find optimized hills. Describe why preservation in the chart was scaling with sequence length.

I don’t have to do ANY of that to show that you have failed to support your claims that the size of sequence space tells us how densely it is packed with functional sequences, or the extend to which different functions overlap in that space.

How well I can support a case for the evolutoinary origin of some protein is completely irrelevant here and it’s obvious you’re just trying to run away from supporting your claims. You seem to be stuck in this mistaken opinion that we need to prove the truth of some alternative to the one you already believe. That’s not how it works, nobody has to actually chose some position. The question is whether any position has good support, and in the argument we’re having right now we’re analyzing a claim you’ve made to see if you actually have any good reason for taking that position in the first place. Whether I can defend my own position, and whether I can persuade you of it, is besides the point.

112

It doesn’t really matter for this discussion. The claim being analyzed here is yours, which is that because we can calculate the total size of sequence space from the length of some protein, we can also know how frequently functions can be found in that entire space.

And when asked how you accomplish this, you invoke the degree to which the Prp8 protein is conserved between a handful of species and argue this means other sequences are selected against. We agree with you they’re selected against, but we want to know how you determine whether they’re selected against because they’re lethal, or because they just have lower fitness.

I just ran blast on brr2 for humans mice and rats and it is 99.5% similar so it also found the theoretical peak in almost infinite search space.

How do you know it’s the “theoretical peak” in that entire landscape? How could you possibly know that merely from how similar the human protein is to the rat version? Explain how you derive that conclusion.

114

The answer is no however the data I showed this morning shows mammals may fit this description.

Evolution can search around 150 bits. Prp8 is 4000 bits. There is a huge gap here to close for evolutionary mechanisms to explain this.

From the mouse human data it looks like purifying selection from an optimized sequence that is the product of a mind.

A high level of sequence preservation points to lethal or you would expect more variation over millions of years.

As a minimum you need to be able to splice out introns.

Again the level of conservation.

It cannot be slightly worse and support mammals.

I am saying what we are observing does not appear to look like a hill because of such high levels of preservation.

This is not my claim. The application such as splicing determines if a hill exists.

You are changing the argument. The argument is that unless substitutability of sequences scales with sequence length the current theory fails. The data in gpuccio’s graph does not support this and the gap is a 1000 orders of magnitude in some cases.

115

Please support this claim with evidence from the primary scientific literature. How did you obtain that value and why is that a limit?

116

Tim this is a common accepted calculation. I don’t think Rum will object. 150 bits is about 10^50 which is greater than the estimated number of evolutionary trials.

118

You can define this as the number trials for a selectable step. Selection does not kick in until you have reached some nominal function. Prp8 is not sufficient in itself to perform a eukaryotic function.

121

Why did I know you’d just regurgitate the same IDC bullshit without providing your calculations? :slightly_smiling_face:

Please quit hiding my response. it’s on topic and relevant.

1 Like
122

No one I know accepts it. Please back up the claim, don’t just regurgitate IDiot talking points.

Back up this number with some evidence too. IDiot calculations never take into account the effects of selection feedback in the search so always come out with these complete garbage "it’s too improbable!’ numbers.

Please quit hiding my relevant and on topic posts.

123

Do you understand that selection requires function?

10^50 is all the trials of all the animals who have lived. If you have a better number I will accept yours. The issue is that the gap is so large that changing this number by 10 orders of magnitude does not change the argument.

124

Do you understand you’ve never shown simpler precursors have no function? And ignored all the evidence functional simpler precursors do in fact exist?

Show your calculations for number of “trials” and justify the assumptions you made. You’re still just regurgitating worthless garbage numbers you read on an IDC page somewhere.

125

That number is not too bad an estimate - roughly 10^30 organisms with 10^4 genes each for 10^12 generations.

But it has nothing whatsoever to do with the number of bits on information in a specific protein, since evolution didn’t need to generate every single possible protein of Prp8’s length in order to produce Prp8. Mainly because proteins often evolve by combining two existing proteins into one longer protein.

126

What evidence?