Evolutionist: You’re Misrepresenting Natural Selection

_{Cornelius Hunter

December 25, 2011

Intelligent Design}

Share: Facebook; Twitter; LinkedIn; Flipboard; Print; Email

How could the most complex designs in the universe arise all by themselves? How could biology’s myriad wonders be fueled by random events such as mutations? Read more

Comments

"You have a bulls eye painted after the arrow has landed."
Clearly this is not the case, as has been shown. Proteins which fold do so regardless of whether anyone's "painted a bullseye" around them. Your claim has been refuted. Venn diagram: Draw a rectangle which represents the sample space. In the center of that, draw a circle which represets the set of folding sequences. In the center of that circle, draw another which represents the functional set. There's the target, and it's objectively real. It cannot reasonably be denied. Plain and simple, neat and clean. The target which constitutes folding, functional proteins is carved upon the face of reality by the laws of physics. It doesn't get any more objective than that. Miller's Blundermaterial.infantacy_{December 31, 2011
December
12
Dec
31
31
2011
03:39 AM
3
03
39
AM
PDT}

As already stated, the set of functional sequences is a subset of the set of folding sequences, which is a subset of all ordered sequences.
That's not entirely true. There are sequences that participate in regulatory networks, and there are probably sequences that have unidentified functions. But unless you have a database of functional sequences or a theory of what makes a sequence function, you can't design, except by cut and try. My point is and always has been that in the absence of the ability to distinguish a functional sequence independently of trial and error, design is impossible. As design advocates you really need to demonstrate that design is theoretically possible. You could answer my challenge by simply telling me how to distinguish a functional sequence with one character altered from a batch of randomly generated sequences. If you can't tell that a non-functional sequence has 149 bits of dFSCI, you can't claim the functional sequence has 150 bits. You don't have a metric. You have a bulls eye painted after the arrow has landed. You simply can't wait for the coins to be tossed before defining a sequence as a target. Design advocates assert that designers have some shortcut to knowing that a sequence will result in a useful fold, but they haven't proposed a design process that doesn't use evolution.Petrushka_{December 31, 2011
December
12
Dec
31
31
2011
12:52 AM
12
12
52
AM
PDT}

Correction: "c) n(F) > 1/n(S), that is, F is not the same as some arbitrary sequence in S, n(F) > 1 (because there is more than one sequence which folds)." Should be "c) P(F) > 1/n(S), that is, F is not the same as some arbitrary sequence in S, n(F) > 1 (because there is more than one sequence which folds)."material.infantacy_{December 30, 2011
December
12
Dec
30
30
2011
05:37 PM
5
05
37
PM
PDT}

"I need not explanation or retraction. Every sequence of coin tosses is exactly as probable as every other sequence….You have no independent way of declaring a sequence to be functional, that is before you test it for function or observe it to be functional."
The example I provided at #10 obviates the need to define function in order to arrive at an objective assessment that function is non-arbitrary, and exists in a remote part of protein sequence space. The equality of outcome for any given sequence is entirely irrelevant. Here's another stab at it. As already stated, the set of functional sequences is a subset of the set of folding sequences, which is a subset of all ordered sequences. This implies that for a protein to be functional, it needs to fold -- and since we know that a subset of sequences fold, we have an objective measure that there exists a set of objectively defined sequences, of which some will be functional, and apart from which none will be. Again, the set of functional sequences is a subset of the set of folding sequences, which provides an objective boundary for the number of functional protein folds. S = {all possible sequences of length n} F = {elements in S which fold} F1 = {functional sequences} F1 ⊆ F ⊂ S n(F1) ≤ n(F) so P(F1) ≤ P(F) P(F) = n(F) / n(S) < 1.0 There is a partition of S, such that F ∪ F' = S, and F ∩ F' = {}. (The set F' is the complement of F, or every element of S which is not in F.) This is to reiterate that functional folds are not arbitrary, regardless of whether or not we can determine/define them. Also, n(F) is minuscule in size compared to n(F'). Here's what's at issue: a) not all sequences fold; b) set F1 (functional sequences) is a subset of F (folding sequences), so the size of F1 is bounded by the size of F, n(F1) ≤ n(F), implying that the probability of F1 occurring is less than or equal to the probability of F occurring; c) n(F) > 1/n(S), that is, F is not the same as some arbitrary sequence in S, n(F) > 1 (because there is more than one sequence which folds). d) functional sequences are irrelevant here (see point b) and only folding sequences need be considered for the objective assessment that not all sequences are equal with regard to having function. Since we know that not all sequences are equal -- that is, some of them fold, and of the ones that do some of them can have a function -- it cannot possibly be more irrelevant that any given sequence is just as improbable as any other. As a matter of fact, it couldn't be more relevant that sequence specificity is king with regard to biological function. Again, to suggest otherwise is to imply that in a biological context, one sequence is as good as any other, which is objectively not the case.material.infantacy_{December 30, 2011
December
12
Dec
30
30
2011
04:53 PM
4
04
53
PM
PDT}

Testing ∩ ∪ ⊆ ⊂ ≤ < > xn A1material.infantacy_{December 30, 2011
December
12
Dec
30
30
2011
04:10 PM
4
04
10
PM
PDT}

I need not explanation or retraction. Every sequence of coin tosses is exactly as probable as every other sequence. You seem to be arguing as if certain sequences are predicted or specified in advance, but the fact is your metric depends on a sequence having some consequence or function before you label it specified. You have no independent way of declaring a sequence to be functional, that is before you test it for function or observe it to be functional. You have no theory of why some sequences are functional and others not, no means of predicting the function of sequences. Therefor RMNS and design have the same problem (or require the same characteristics of functional space. If function is truly scattered and sparse, with no connecting ridges, then design and evolution are both impossible. If functional space is connectable, then the design hypothesis is unnecessary.Petrushka_{December 30, 2011
December
12
Dec
30
30
2011
01:51 PM
1
01
51
PM
PDT}

Well, I think it's yours :) But then I would, wouldn't I? Sleep well :)Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
01:42 PM
1
01
42
PM
PDT}

Let’s be more clear. Let’s start from a duplicated gene. You make a big fuss about the gene retaining for some time its original function, the double allele, and so on. None of that is relevant.
Well, it was relevant to the point you originally made with presented only two alternative scenarios, and dismissed both! I presented a third, and defended that!
Let’s say that, if that duplicate gene will become the seed for a new domain, unrelated to the original at sequence level (as it must happen, because new domains do appear), at some point it must lose its original function and become inactive.
And immediately you introduce new cans of worms! The "seed" for "a new domain"? Do you just mean that the duplicated sequence may eventually become the coding sequence for a new gene? What is "seed" doing in there? And why must it become inactive before it does something new? Why could it not simply undergo series of point mutation and produce different a slightly protein at each change? Until one day, the protein produced turns out to have a useful new function function? Or have undergo some change to its promotor so it is now expressed in different tissue, or at a different developmental stage? Or, as you could say, it could become inactive, or active but useless (produce some harmless but non-functional protein).
At this point, all mutations are neutral.
. No they aren't, necessarily. Some may well be deleterious if they result in a toxic protein.
As mutations accrue, its sequence will no more be related to the original sequence. At this point, we can safely say that any unrelated state of the sequence has the same probability to be reached by a random wlak as any other.
I agree that the longer it remains functionless, the further it will move from its "parent" sequence. It's destiny may well be a pseudogene.
Your reasoning that “if it was originally a protein-coding gene, it’s not going to go far from being a protein coding gene” has no sense. Any stop codon, or frameshift mutation, will make an ORF no more an ORF.
That's true. But back mutations are also possible. Also restorations of the reading frame. And some frameshifts result in viable proteins.
And even if the oRF is still transcribed or translated, the protein will no more fold or be functional.
Most probably. But not necessarily. Brand new genes are probably relatively rare (compared with new alleles, or with duplications). But that doesn't mean they won't ever happen. Mutations are common.
We will have a pseudogene, or a protein coding gene that codes for a completely non functional protein (a protein is a protein even if it does not fold, even if it has no function).
Sometimes indeed this will happen. More often then not, in fact. Hence the notorious "junk DNA". But "rarely" is a a lot different from "never", especially given so many opportunities.
So, new domains are exactly that: some sequence pulled nucleotide by nucleotide from a black bag.
No. I might even run a simulation to show you that this is false. The variance you'd get by selecting nucleotide by nucleotide from a black bag is orders of magnitude higher than the variance you'd get if you started with a certain sequence, mutated it in various ways known to happen, and weeded out those turned out to be lethal.
Or rather, some sequence that has to be reached by a completely blind random walk from a completely unrelated point of the search space. (Indeed, the idea of the black bag is incorrect, because it is related to a random search by successive extractions, and not to a random walk. The correct model is a random walk).
Yes, the correct model is a random model. Good, I don't have to write the simulation! Or perhaps I do. I'll give it a go, tomorrow.Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
01:40 PM
1
01
40
PM
PDT}

Unicellular organisms have a body and therefor they also have a body plan. HOX genes require an explanation before you can use them for any explanations.Joe_{December 30, 2011
December
12
Dec
30
30
2011
01:35 PM
1
01
35
PM
PDT}

Elizabeth: Frankly, I don't think it is a problem of language. I think it is a cognitive problem. With respect, your copgnitive problem. What can I say? I disagree with all, but I am tired. Maybe tomorrow...gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
01:33 PM
1
01
33
PM
PDT}

Petrushka: I have done the discussion about the paths thousands of times. I cannot always repeat everything. The above discussion was about the random part. Please, reread all my post with honesty, and don't change the discussion. We are still waiting for your answer about probabilities, and for a defense, or a retraction, of your statement in post 2.3.2.1.14.gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
01:24 PM
1
01
24
PM
PDT}

Elizabeth: So I’d say that all systems are stochastic. So it doesn’t tell us much in itself ? No. I would not say so. Conventional physics is deterministic. You have taken an example from quantum physics, which is both deterministic and probabilistic, at different points. In conventional physics, the tossing of a coin is deterministic, but as we cannot know all the variables implied, we can better describe it with a probabilistic system, and if we assune it is a fair coin, a discrete uniform distribution, with each of the two events having a 0.5 probability, will describe the system quite well. Even in quantum mechanics, the evolution of the wave funcion is strictly deterministic. It’s only the “collapse” of the wave function that is probabilistic, and I must say that the meaning of probability in the qunatum world is controversial: it could be intrinsic randomness, or just a pseudo-randomness like in classical random systems, although I suppose that the first position prevails.
You put your finger on it IMO when you say “as we cannot know all the variables implied”. Yes indeed. And we never do. Even if at some fundamental level the universe proves to have no quantum uncertainty (as at least one eminent theoretical physicist has proposed) all our models (and we are talking about models here) must be stochastic. Now physicists work to find tolerances and insist on 5 sigma confidence, while life scientists are often content with two. But every single effect we observe comes with stochastic variance. We can never know all the variables. What we can do is estimate probability distributions, and also the extent to which those distributions are orthogonal.
Well, as I said, I’d say it’s all stochastic, and in no case are we dealing with flat probability distributions. Mutations have associated non-flat probability distributions, as does the probability that a variant with given properties will propagate through a population. Confusion, again. I suppose that with “flat” you mean a uniform probability distribution. But there is no need that the probability distribution that describes a system should be flat. It can be any valid probability function. The system is random all the same.
I agree. That was my point. All systems are stochastic. What we need are the relevant probability distributions, not a division into stochastic and non-stochastic processes. They are all stochastic. Some merely have more variance than others. This is, btw, why I avoid the term “RM+NS” and get cross when people say that mutation is random and natural selection isn’t. Both are variance generation (“RM”) and differential reproduction (“NS”) stochastic processes. And both have probability distributions that are biased in favour of reproductive success.
That mutations can have slightly different probabilities is true. But I will try to explain, when we arrive to that point, why that is not really relevant.
OK.
Instead, your observation about the effect of NS in propagating a mutation (I suppose you are referring to NS, as you speak of “properties”) is a necessity effect. That has nothing to do with the random system. It is a necessity algorithm that is “coupled” to the random system that generates RV. More on that later (I hope).
Well, as I keep saying, I think these terms and categories are misleading you. Darwinian evolution is extremely simple: if organisms reproduce with heritable variation in reproductive success, variants that tend to reproduce more successfully will become more prevalent. That’s the algorithm. It’s not “coupled” to anything. What is “coupled” within the algorithm, is reproduction with heritable variation in reproductive success. Take away either half of the “coupling” and you don’t have an algorithm. I honestly think that much of the flak copped by Darwinism boils down to overcomplication. That’s all it is. What is genuinely complicated of course are the mechanisms by which variation is generated and replicated. And those are fascinating, and many are still a mystery. But the algorithm is dead simple.
I do not find that link at all enlightening. Please give me an example of a function that is not “naturally selectable”. Most biochemical functions are not “naturally selectable” in a specific context. Supppose I introduce by genetic engineering a gene that codes for some human protein into a bacterial population, where the biochemical context that uses the biochemical function of the gene is not present. In that case, the simple existence of a protein that, while retaining its biochemical function, is not in the context to use it, will not add to the reproductive potential of the bacteria. It will not be naturally selectable.
Notice that you could only give me an example from Intelligent Design :) Try again please :) This is important.
To be naturally selectable, a protein gene must code for a protein that can be usefully integrated in the existing biochemical environment of the species where it emerges.
Sure. Because if it codes for a protein that performs no function for the phenotype, and can perform no function for the phenotype, then it doesn’t have a function, does it? Sure a protein might be generated and not do anything for the organism. But then it would be a functionless protein wouldn’t it? How can it have a function and not do anything useful? Isn’t that an oxymoron? Does it work in Italian?
Genes do not “implement a function”. Hey, that’s really sticking to words and being fastidious.
Yup. I know. The devil is in the details. I’m trying to get past those waving hands (Granville’s too) and these high level metaphors and personification (NS as agent) and down to the nitty gritty of what actually happens!
A protein coding gene codes for a protein. The protein implements a function.
No, the protein doesn’t “implement a function”. The protein may play a role in some function but that doesn’t mean it “implements” it. These things matter. A protein produced in the wrong place and/or the wrong time may cause serious disease. The function is much larger than the protein. At least it is if we are talking about multicellular organisms. Maybe not so much if we are talking about unicellular organisms (which are not my speciality :))
OK, we are writing on a blog, in a hurry. Could you please try to understand what I mean? The gene has the information for the protein. The protein imnplements a function. I have no confusion about genotype and phenotype.
I know you think you don’t, and I know you think I am being difficult. But from where I’m standing you do have a problem, and I think it’s because you have over-simplified to the point of falsification the role that proteins play in a functioning organism. I’ve recommended this video many times, I don’t know whether you have seen it. I think it’s excellent (and a refreshing antidote to Dawkins’ “Selfish Gene” concept): http://videolectures.net/eccs07_noble_psb/ I do wish people on this blog would watch it and take note. It would actually improve some of the arguments both for and against Darwinian evolution!
They are part of a large system of genes that work together to produce a co-ordinated functional phenotype. What does that mean? The glucose 6 phosphatase gene codes for a specific enzyme, that has a specific biological function. That such a function is integrated in a higher level organization is true, but that does not make the local function less real. And you cannot integrate functions that do not exist into an higher level organization. Your reasoning really makes no sense.
Yes it does :) I didn’t say that functions couldn’t be local. Indeed it’s important that they are. If you start producing enzymes in the wrong tissue at the wrong time, you produce disorders. But if the local system doesn’t help the phenotype (the whole organism) reproduce then it’s not fulfilling a function, whatever else it does. Not in the normal sense of the word “function” anyway. It would be what my husband calls our cats: “do-nothing-machines”.
If your darwinist education prevents you, an intelligent person, from understanding that the local function of an enzyme is different from the global function of an integrated protein system, that still is different from the general reproductive function of a living being, then I must say that darwinism is really much more dangerous for human mind than I believed
Well, it doesn’t. So retract that thought.
Most genes are expressed in many organs and tissues, and, depending on where and when they are expressed, serve a different function, No. Their biochemical activity remains the same.
I didn’t say it didn’t. They nonetheless may serve different functions, i.e. some genes are pleiotropic. Even within one organ, a single gene, for example a neurotransmitter transporter, may serve different functions depending on where, how, and the degree to which it is expressed.
Unless a gene does something that contributes to the reproductive success of the organisms of a population I don’t see how you can say it has a function. Hey! What are you saying? An enzyme retains its biochemical function in the lab, in a cell free system, where it does not certainly contribute to any reproductive success!
Well, it retains its biochemical properties – it still catalyses the same reactions. I wouldn’t call that a function. I’d call it a biochemical property.
DNA polymerase is standardly used in labs, because it does what it does. I don’t think that allowing us to perform PCR is contributing to some reproductive success. And yet, we can perform that technique in the lab because the involved enzymes retain their biochemical functions even if completely separated from the living beings where they were formed, and even if artificially built in the lab in a cell free system. How can you make such senseless statements?
It could actually be a language problem. I was helping my son translate some Spanish poetry this evening, and we were both commenting that one Spanish word can have many English equivalents, and so you lose in translation the multivalence of the Spanish word. I’ve found the same with Italian too. English is wonderful for precision, but romance languages can be better for philosophy! So no, my statement is not senseless. But if you want to use the word “function” to refer to what I would call a simply a chemical or physical “property”, then it is important that we keep those meanings separate and do not equivocate between the two. By “function” I mean: serves some purpose, whether teleologic or teleonomic. In the case of components of living things, I mean teleonomic purpose – the role a component plays in contributing to the reproductive success (including the maintenance) of the organism. I might also use an enzyme teleologically to help me get my clothes clean, but or even as some kind of reagent in a lab, in which case I would be giving it a function. But just sitting there catalysing for no purpose, that, I wouldn’t call having a “function” merely exhibiting a biochemical property.
Well, they seem unnecessarily complicated and potentially misleading. For a start, “natural selection” occurs at the level of the phenotype, not the genotype (as I keep saying, you seem to confuse these levels). I am nbot confusing anything. NS does happen at the phenotype level, but as its result is differential reproduction, its relevant effect is expanding or reducing the instances of a specific genome. That’s the only thing that is relevant to evoultion. So, the genetic variation gives a phenotypic effect, the phenotypic effect determines the expansion or contarction of the population, and therefore the expansion or contraction of the new gene. And so? Must I write all that each time to make you happy? What changes?
OK. I wasn’t sure what you meant by “expansion”. I’d call that “propagation” or “replication” :)
You cannot consider natural selection at the molecular level only. Why? I consider the effects of NS at the molecular level, because that is the level where the information is, and it’s information we are debating. Why in th world cannot I do that?
Because you need also to consider the phenotype!!!!! The phenotype does not even exist at the molecular level!
C1) A neutral variant is one that either confers no greater reproductive success on its bearer than the parental variant did on its bearer in the same environment or that confers no greater reproductive success on its bearer in the current environment than the mean reproductive success of all other variants in the population. Simple and intuitive indeed! Look at mine: c1) Non visible to NS: that variation in no way modifies the reproductive potential of the reproducer compared to the rest of the population. That’s what a “neutral” variation is. Where is your problem? The variant is different: either that difference changes something in the reproductive power of the varied being, or not. And your complex, and useless, definition is misleading to our discussion. After all, we are looking for how new functional proteins emerge (do you remember? that was the scenario). If a modified gene (and protein sequence, just to be complete) does not code for a protein that is new and functional, it can be in some clone that expands or not, for other reasons linked to other parts of the genome, or just because of drift: all that si possible, but in no way that will select for molecular function. In no way it will help a new functional protein to emerge.
Oh boy. My definition is not complex at all. It doesn’t require us to imagine NS as having metaphorical “eyes” to which “variation” might be “visible”. It just tells you what the word means. Let’s leave the rest for now:
You are always at the same point: you imagine that functional proteins can emerge by chance alaone, and still you want to keep NS in the field, just to be sure. Well, your reasoning is faulty and wrong.
And here we are right back where we started with “chance alone”. Look, my position, as I’ve said, is that the entire system is stochastic (i.e. has “chance” elements) both variation and “NS”. Of course I want “NS in the field”!!! It’s sitting there already!!! I do understand that you think my “reasoning is faulty and wrong” but from where I’m standing you seem not to have understood any part of it! Perhaps, as I said, that is a language problem. I will sleep on this and see if I can be clearer. Communication is hard!
But bear in mind that if the environment changes, that variant may become either beneficial or deleterious. Most new variants are near-neutral when they first appear. Bear also in mind that allele frequency itself is a powerful element in the environment. The simple reason why you run away from the molecular scenario is that, at that level, nothing of what you say has any sense.
I’m not “run[ning] away from the molecular scenario”. I honestly have no clue why you think so. What I am insisting on is that when we evaluate the function specified by DNA (a molecule) we need to do so in terms of its effect on the phenotype. You keep saying you are remembering the phenotype, then you ignore it!
Again, proteins must fold well, must have an active site, must have a specific, oftem amazing, biochemical activity. Otherwise they are useless.
And even if they have that biochemical activity, they can be useless just the same in the wrong context. And so? Penicillinase is not useful if there is no penicillin in the environment. But it still is a wonderful biochemical machine. Sure. But has no function, in my usage. It may have very special biochemical properties.
But if you have to metabolize nylon, you use the same penicillinase strucure, with minimal modifications, because it’s always an esterase activity you need. If you had not penicillinase, nylonase could never emerge. Except in Ono’s imagination…
Exactly. A functional protein can have a non-functional precursor. That’s what I’ve been saying all along. And that non-functional precursor might exist because it was, in the past functional.Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
01:20 PM
1
01
20
PM
PDT}

So, new domains are exactly that: some sequence pulled nucleotide by nucleotide from a black bag. Or rather, some sequence that has to be reached by a completely blind random walk from a completely unrelated point of the search space.
You are simply assuming that there is no historical path leading to modern domain sequences. You've done no research to demonstrate the lack of a path. You have no theory that speaks to the question of whether there is a path. All you have are sequences that have no living cousins. That is not positive evidence to support an unobserved designer.Petrushka_{December 30, 2011
December
12
Dec
30
30
2011
01:03 PM
1
01
03
PM
PDT}

Elizabeth: Just a comment on this statement fron you: No. Again, you forget what is given. The duplicated gene is certainly not “free to go anywhere” – it’s highly constrained by what it is when it stops being what it was. In other words, if it was originally a protein-coding gene, it’s not going to go far from being a protein coding gene. And “protein coding space” is highly clustered – any changes to an ex-protein coding sequence is far more likely to hit a new protein coding sequence than, for example, some sequence pulled nucleotide by nucleotide from a black bag. And that’s what we see – genes in “families” where the gene-lineage is apparent in the sequence. Again, what do you mean? My discussion is about the 2000 protein superfamilies that are not clustered at all. It's obvious that, in a single family or superfamily, there are similatiries. That's why I don't usually discuss the evolution inside families (that can, anyway, be discussed). I usually admit that the evolution inside families is of two kind: a) With preservation of more or less the same function, and variation of the primary sequence out of neutral variation. b) With slight, sometimes more radicval, "tweaking" of the function, and levels of variation that could be borderline for a design detection (5 - 10 aminoacids). But the true ID argument is about the fundamental islands, the basic domains. Those are not clustered. They are completely isolated at primary sequence level (less than 10% homology), have different folding and different function. That's why I say that, in that context, each unrelated state has more or less the same probabilities, and therefore we can assume a practically uniform distribution. Let's be more clear. Let's start from a duplicated gene. You make a big fuss about the gene retaining for some time its original function, the double allele, and so on. None of that is relevant. Let's say that, if that duplicate gene will become the seed for a new domain, unrelated to the original at sequence level (as it must happen, because new domains do appear), at some point it must lose its original function and become inactive. At this point, all mutations are neutral. As mutations accrue, its sequence will no more be related to the original sequence. At this point, we can safely say that any unrelated state of the sequence has the same probability to be reached by a random wlak as any other. Your reasoning that "if it was originally a protein-coding gene, it’s not going to go far from being a protein coding gene" has no sense. Any stop codon, or frameshift mutation, will make an ORF no more an ORF. And even if the oRF is still transcribed or translated, the protein will no more fold or be functional. We will have a pseudogene, or a protein coding gene that codes for a completely non functional protein (a protein is a protein even if it does not fold, even if it has no function). So, new domains are exactly that: some sequence pulled nucleotide by nucleotide from a black bag. Or rather, some sequence that has to be reached by a completely blind random walk from a completely unrelated point of the search space. (Indeed, the idea of the black bag is incorrect, because it is related to a random search by successive extractions, and not to a random walk. The correct model is a random walk).gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
12:54 PM
12
12
54
PM
PDT}

Elizabeth: So I’d say that all systems are stochastic. So it doesn’t tell us much in itself ? No. I would not say so. Conventional physics is deterministic. You have taken an example from quantum physics, which is both deterministic and probabilistic, at different points. In conventional physics, the tossing of a coin is deterministic, but as we cannot know all the variables implied, we can better describe it with a probabilistic system, and if we assune it is a fair coin, a discrete uniform distribution, with each of the two events having a 0.5 probability, will describe the system quite well. Even in quantum mechanics, the evolution of the wave funcion is strictly deterministic. It's only the "collapse" of the wave function that is probabilistic, and I must say that the meaning of probability in the qunatum world is controversial: it could be intrinsic randomness, or just a pseudo-randomness like in classical random systems, although I suppose that the first position prevails. Well, as I said, I’d say it’s all stochastic, and in no case are we dealing with flat probability distributions. Mutations have associated non-flat probability distributions, as does the probability that a variant with given properties will propagate through a population. Confusion, again. I suppose that with "flat" you mean a uniform probability distribution. But there is no need that the probability distribution that describes a system should be flat. It can be any valid probability function. The system is random all the same. That mutations can have slightly different probabilities is true. But I will try to explain, when we arrive to that point, why that is not really relevant. Instead, your observation about the effect of NS in propagating a mutation (I suppose you are referring to NS, as you speak of "properties") is a necessity effect. That has nothing to do with the random system. It is a necessity algorithm that is "coupled" to the random system that generates RV. More on that later (I hope). I do not find that link at all enlightening. Please give me an example of a function that is not “naturally selectable”. Most biochemical functions are not "naturally selectable" in a specific context. Supppose I introduce by genetic engineering a gene that codes for some human protein into a bacterial population, where the biochemical context that uses the biochemical function of the gene is not present. In that case, the simple existence of a protein that, while retaining its biochemical function, is not in the context to use it, will not add to the reproductive potential of the bacteria. It will not be naturally selectable. To be naturally selectable, a protein gene must code for a protein that can be usefully integrated in the existing biochemical environment of the species where it emerges. Genes do not “implement a function”. Hey, that's really sticking to words and being fastidious. A protein coding gene codes for a protein. The protein implements a function. OK, we are writing on a blog, in a hurry. Could you please try to understand what I mean? The gene has the information for the protein. The protein imnplements a function. I have no confusion about genotype and phenotype. They are part of a large system of genes that work together to produce a co-ordinated functional phenotype. What does that mean? The glucose 6 phosphatase gene codes for a specific enzyme, that has a specific biological function. That such a function is integrated in a higher level organization is true, but that does not make the local function less real. And you cannot integrate functions that do not exist into an higher level organization. Your reasoning really makes no sense. If your darwinist education prevents you, an intelligent person, from understanding that the local function of an enzyme is different from the global function of an integrated protein system, that still is different from the general reproductive function of a living being, then I must say that darwinism is really much more dangerous for human mind than I believed :) Most genes are expressed in many organs and tissues, and, depending on where and when they are expressed, serve a different function, No. Their biochemical activity remains the same. Unless a gene does something that contributes to the reproductive success of the organisms of a population I don’t see how you can say it has a function. Hey! What are you saying? An enzyme retains its biochemical function in the lab, in a cell free system, where it does not certainly contribute to any reproductive success! DNA polymerase is standardly used in labs, because it does what it does. I don't think that allowing us to perform PCR is contributing to some reproductive success. And yet, we can perform that technique in the lab because the involved enzymes retain their biochemical functions even if completely separated from the living beings where they were formed, and even if artificially built in the lab in a cell free system. How can you make such senseless statements? Well, they seem unnecessarily complicated and potentially misleading. For a start, “natural selection” occurs at the level of the phenotype, not the genotype (as I keep saying, you seem to confuse these levels). I am nbot confusing anything. NS does happen at the phenotype level, but as its result is differential reproduction, its relevant effect is expanding or reducing the instances of a specific genome. That's the only thing that is relevant to evoultion. So, the genetic variation gives a phenotypic effect, the phenotypic effect determines the expansion or contarction of the population, and therefore the expansion or contraction of the new gene. And so? Must I write all that each time to make you happy? What changes? You cannot consider natural selection at the molecular level only. Why? I consider the effects of NS at the molecular level, because that is the level where the information is, and it's information we are debating. Why in th world cannot I do that? C1) A neutral variant is one that either confers no greater reproductive success on its bearer than the parental variant did on its bearer in the same environment or that confers no greater reproductive success on its bearer in the current environment than the mean reproductive success of all other variants in the population. Simple and intuitive indeed! Look at mine: c1) Non visible to NS: that variation in no way modifies the reproductive potential of the reproducer compared to the rest of the population. That’s what a “neutral” variation is. Where is your problem? The variant is different: either that difference changes something in the reproductive power of the varied being, or not. And your complex, and useless, definition is misleading to our discussion. After all, we are looking for how new functional proteins emerge (do you remember? that was the scenario). If a modified gene (and protein sequence, just to be complete) does not code for a protein that is new and functional, it can be in some clone that expands or not, for other reasons linked to other parts of the genome, or just because of drift: all that si possible, but in no way that will select for molecular function. In no way it will help a new functional protein to emerge. You are always at the same point: you imagine that functional proteins can emerge by chance alaone, and still you want to keep NS in the field, just to be sure. Well, your reasoning is faulty and wrong. But bear in mind that if the environment changes, that variant may become either beneficial or deleterious. Most new variants are near-neutral when they first appear. Bear also in mind that allele frequency itself is a powerful element in the environment. The simple reason why you run away from the molecular scenario is that, at that level, nothing of what you say has any sense. Again, proteins must fold well, must have an active site, must have a specific, oftem amazing, biochemical activity. Otherwise they are useless. And even if they have that biochemical activity, they can be useless just the same in the wrong context. And so? Penicillinase is not useful if there is no penicillin in the environment. But it still is a wonderful biochemical machine. But if you have to metabolize nylon, you use the same penicillinase strucure, with minimal modifications, because it's always an esterase activity you need. If you had not penicillinase, nylonase could never emerge. Except in Ono's imagination...gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
11:40 AM
11
11
40
AM
PDT}

Elizabeth: Nice to talk you again, for me too! And my best wishes of a happy New Year. I don’t think I have the time now to answer all your points. IMO, the confusion of your arguments is huge, but I am sure you think the same of mine I will just start by clarifying some of my terms which seem to give you problems. a) A “random system” is a system whose behaviour cannot be described in terms of necessity laws, usually because we cannot know all the variables and/or all the laws, but which can still be described well enough in terms of some probability model. The tossing of a coin is a random system.
OK, thanks. That’s what I would call a “stochastic” system, but I’m happy to use your term. However I have issues about those “necessity laws”. It seems to me that “necessity laws” are also stochastic, and what we have instead is a continuum of uncertainty. For instance we can radioactive decay is, at one level, a “necessity law” and we can make highly precise predictions about half-lives etc at the macro level, but very poor prediction at the level of individual decay events, at which point we are reduced to a probability distribution as with any stochastic system. So I’d say that all systems are stochastic. So it doesn’t tell us much in itself ?
Genetic variation, if we don’t consider the effects of NS (differential reproduction) is a random system. NS introduces an element of necessity, due to the interaction of reproductiove functions and the environment.
Well, as I said, I’d say it’s all stochastic, and in no case are we dealing with flat probability distributions. Mutations have associated non-flat probability distributions, as does the probability that a variant with given properties will propagate through a population.
b)”Function” and “naturally selectable function” are two different things. Please, look at my answers to englishmaninistanbul (and Petrushka) here: https://uncommondescent.com.....ent-412908 for my definition of “function”. The biochemical action of an enzyme is its function.
I do not find that link at all enlightening. Please give me an example of a function that is not “naturally selectable”.
A “naturally selectable function” is a biochemical function that, in a specific biological being, confers a reproductive advantage. IOWs, if you add the gene that implements that function, the reproducer with that gene will reproduce better, and the prevalence of the gene in the population will be amplified: that’s what I mean by “expansion”.
The reason this makes no sense, and, as a result, your distinction between “function” and “naturally selectable function” is a distinction without a difference, IMO, is that you have, again, confused genotype with phenotype! Genes do not “implement a function”. They are part of a large system of genes that work together to produce a co-ordinated functional phenotype. Most genes are expressed in many organs and tissues, and, depending on where and when they are expressed, serve a different function, but none are solely responsible for that function. Perhaps the distinction you are trying to get at is between a gene that is sometimes expressed as a protein (i.e. the DNA has the “function” of sometimes resulting in a protein) and between a gene that makes the difference between a successfully reproducting organism and an unsuccessful one. But I think that’s a very poor use of terms myself. Unless a gene does something that contributes to the reproductive success of the organisms of a population I don’t see how you can say it has a function. And if it does, then it’s by definition selected. That would include, of course, any gene that contributes to the maintenance of the organism.
c) For NS I just mean what darwinists mean, only I refer always to the molecular aspect. So, a molecular variation can be: c1) Non visible to NS: that variation in no way modifies the reproductive potential of the reproducer compared to the rest of the population. That’s what a “neutral” variation is. c2) Visible to negative selection: The bearer of the variation will reproduce worse compared to the rest of the population. The vairation will tend, more or less, to be eliminated from the population genome. c3) Visible to positive selection: the bearer of the variatio reproduces better, and will expand in the population (or, which is the same, those who do not have the new trait will reproduce relatively worse. The new trait, conferring the new biological function, for instance a new enzymatic activity, will expand in the population genome. I can’t see what is the problem with those concepts. Could you please confirm if you accept them, and if not, why? Otyherwise, it’s hopeless to go on…
Well, they seem unnecessarily complicated and potentially misleading. For a start, “natural selection” occurs at the level of the phenotype, not the genotype (as I keep saying, you seem to confuse these levels). Natural selection is, to a “Darwinist”, simply heritable variation in reproductive success. You cannot consider natural selection at the molecular level only. And in place of your c subdivisions, I would say: C1) A neutral variant is one that either confers no greater reproductive success on its bearer than the parental variant did on its bearer in the same environment or that confers no greater reproductive success on its bearer in the current environment than the mean reproductive success of all other variants in the population. C2) A deleterious variant is one that confers reduced reproductive success etc. C3) A beneficial variant is one that confers increased reproductive success etc. But bear in mind that if the environment changes, that variant may become either beneficial or deleterious. Most new variants are near-neutral when they first appear. Bear also in mind that allele frequency itself is a powerful element in the environment. Would you agree?Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
09:42 AM
9
09
42
AM
PDT}

Elizabeth: Nice to talk you again, for me too! And my best wishes of a happy New Year. I don't think I have the time now to answer all your points. IMO, the confusion of your arguments is huge, but I am sure you think the same of mine :) I will just start by clarifying some of my terms which seem to give you problems. a) A "random system" is a system whose behaviour cannot be described in terms of necessity laws, usually because we cannot know all the variables and/or all the laws, but which can still be described well enough in terms of some probability model. The tossing of a coin is a random system. Genetic variation, if we don't consider the effects of NS (differential reproduction) is a random system. NS introduces an element of necessity, due to the interaction of reproductiove functions and the environment. b)"Function" and "naturally selectable function" are two different things. Please, look at my answers to englishmaninistanbul (and Petrushka) here: https://uncommondescent.com/intelligent-design/There-is-no-brilliance-in-mechanism-and-reductionism-any-more/comment-page-1/#comment-412908 for my definition of "function". The biochemical action of an enzyme is its function. A "naturally selectable function" is a biochemical function that, in a specific biological being, confers a reproductive advantage. IOWs, if you add the gene that implements that function, the reproducer with that gene will reproduce better, and the prevalence of the gene in the population will be amplified: that's what I mean by "expansion". c) For NS I just mean what darwinists mean, only I refer always to the molecular aspect. So, a molecular variation can be: c1) Non visible to NS: that variation in no way modifies the reproductive potential of the reproducer compared to the rest of the population. That's what a "neutral" variation is. c2) Visible to negative selection: The bearer of the variation will reproduce worse compared to the rest of the population. The vairation will tend, more or less, to be eliminated from the population genome. c3) Visible to positive selection: the bearer of the variatio reproduces better, and will expand in the population (or, which is the same, those who do not have the new trait will reproduce relatively worse. The new trait, conferring the new biological function, for instance a new enzymatic activity, will expand in the population genome. I can't see what is the problem with those concepts. Could you please confirm if you accept them, and if not, why? Otyherwise, it's hopeless to go on...gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
09:10 AM
9
09
10
AM
PDT}

You should know very well that finding a sequence that is part of some specific, and extremely unlikely, subset of events is not pointless at all.
You are arguing as if the sequence is specified in advance. I've asked you repeatedly how you know in advance what a sequence will do. Unless you can demonstrate otherwise, I don't think you can tell if a sequence is random or just one character change from being functional. If you are going to claim that design is possible you need to have a theory of design, a process other than evolution that makes finding functional sequences possible.Petrushka_{December 30, 2011
December
12
Dec
30
30
2011
09:05 AM
9
09
05
AM
PDT}

It was his holocaust denial. One of his many trolls.Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
08:57 AM
8
08
57
AM
PDT}

Bozorghmehr posted here for a few days and was banned. He seems to have made a few converts, though. My memory could be faulty, but I think it was his conspiracy theories that got him banned.Petrushka_{December 30, 2011
December
12
Dec
30
30
2011
08:40 AM
8
08
40
AM
PDT}

tjguy: The paper is exactly right. Let’s just comment very simply the problems of the duplicated gene mechanism as a neodarwinian tool. It’s complete nonsense. The simple truth is that NS acts as negative selection to keep the already existing information. We see the results of that everywhere in the proteome: the same function is maintained in time and in different species, even if the primary sequence can vary in time because of neutral variation. So, negative NS conserves the existing function, and allow only neutral or quasi neutral variation. In that sense it works against any emergence of completely new information from the existing one, even if it can tolerate some limites “tweaking” of what already exists (microevolution).
Again, your metaphor of NS “acting” is misleading you. Natural selection is an effect, not an agent. It doesn’t “act... as negative selection to keep the already existing information”. It doesn’t “act” at all. It is simply the phenomenon by which DNA sequences that only result in reproductive success if unaltered (or only slightly altered) are highly conserved. In other words it’s a way of describing sequences that are both vital and “brittle” – most variants are lethal. It doesn’t “work against the emergence of completely new information”, it simply means that certain sequences are unlikely to be propagated in mutated form. That doesn’t mean the sequence can’t be duplicated, or other less brittle or less vital sequences mutated and propagaged, so it does not have any implications at all for the “emergence of completely new information”.
I suppose that darwinists, or at least some of them, are aware of that difficulty as soon as one tries to explain completely new information, such as a new basic protein domain. Not only the darwinian theory cannot explain it, it really works against it. So, the duplicated gene mechanism is invoked.
Not, invoked, gpuccio, observed, as one of several mechanisms by which new genes, as opposed to new alleles, are created.
The problem is that the duplicated gene, to be free to vary and to leave the original functional island, must be no more translated and no more functional. Indeed, that happens very early in the history of a duplicated gene, because many forma of variation will completely inactivate it as a functional ORF, as we can see all the time with pseudogenes.
Not at all. It can be translated and be functional until it undergoes a mutation that renders it non-functional (and become as you say, a pseudogene). It doesn’t have to become non-function in order to vary. Indeed, one point that Atheistoclast does make, which is quite an interesting one, although he doesn’t make it very clearly, and I’m not even sure he ever really got his own point, is that to some extent, the duplicated gene will sometimes tend to be conserved as a “spare”, as he says, although I think that is a very misleading way of putting it, and indeed, misled him (not that that is very difficult). If the duplicated allele is a useful, but not especially common one, in a sexually reproducing population, individuals with two copies of it may have more offspring than individuals with one, because they are more likely to end up with offspring with at least one. In fact, on a TR thread in which Atheistoclast was developing this idea, he and I both made simple computer models (his was appalling) and I did in fact show that this was true, and so an increasing proportion of the population end up with both the new gene and the old, both still functioning. However, once the prevalence of both in the population gets very high, the reproductive advantage to having a duplicate starts to drop (because your chance of meeting a mate who also has at least one copy starts to rise, so you don’t “need the spare”), and so any inactivating mutation in one of the copies ceases to come with a drop in reproductive success. And once inactivated, it is now “free to mutate”, with the concomitant possibility that one of those subsequent mutations will yield a functional gene.
So, one of the two: a) either the duplicated gene remains functional and contributes to the reproduction, so that negative NS can preserve it. In that case, it cannot “move” to new unrelated forms of function. b) or the duplicated gene immediately becomes non functional, and is free to vary. The important point is that case a) is completely useless to the darwinian explanation. Case b) allows free transitions, but they are no more visible to NS, at least not until a new functional ORF (with the necessary regulatory sites) is generated. IOWs, all variation from that point on becomes neutral by definition.
And you are back with your agency language again! “Visible to NS” is meaningless in this context (if ever meaningful in any context). In any case you’ve excluded the most likely middle which is, as I’ve explained, that the duplicates are conserved until most of the population have at least one copy of the good allele, and two copies of the gene, at which point, having two good copies of the good allele ceases to be strongly advantageous, if advantageous at all, and the copy (or the original, it doesn’t matter) is “freed” from conserving forces i.e. the phenotypes suffer no drop in reproductive success if the sequence suffers a deactivating mutation. Binary thinking seems to be the problem here. Also absolutist thinking. “Neutrality”, “advantageousness”, “deleteriousness” are not absolute terms, they are relative, and they change over time. Evolution is a dynamic process, and the prevalence of alleles in the population is itself part of the environment that determines fitness, so there are powerful feedback loops, as I described above (in my model). And once conserving forces towards having two copies are relaxed, one is free to mutate into something else that may come in handy.
But neutral variation, while free of going anywhere, is indeed free of going anywhere. That means: feedom is accompanied by the huge rising of the probability barriers. As we know, finding a new protein domain by chance alone is exactly what ID has shown to be empirically impossible.
No. Again, you forget what is given. The duplicated gene is certainly not “free to go anywhere” – it’s highly constrained by what it is when it stops being what it was. In other words, if it was originally a protein-coding gene, it’s not going to go far from being a protein coding gene. And “protein coding space” is highly clustered – any changes to an ex-protein coding sequence is far more likely to hit a new protein coding sequence than, for example, some sequence pulled nucleotide by nucleotide from a black bag. And that’s what we see – genes in “families” where the gene-lineage is apparent in the sequence.
IOWs, the neo darwinian “explanation” is silly and wrong.
Well, Bozorgmehr is both silly and wrong. I don’t think you are silly. But I do think you are wrong :)Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
08:38 AM
8
08
38
AM
PDT}

F/N: I think this thread has in it some key discussions, so I have taken liberty to clip and highlight, here. Pardon delay in notification, the struggle with the black screen of sudden death is not over yet -- and Linux is looking better and better to me, 2nd time it has saved my neck. KFkairosfocus_{December 30, 2011
December
12
Dec
30
30
2011
08:11 AM
8
08
11
AM
PDT}

Elizabeth: If you remembered the definition of dFSCI, that I have patiently discussed with you many times. you would know that dFSCI refers to the total information necessary to get a function.
Get to a function from where?
If your 1 bit increases are functional and selectable, then show them. If they are not, then the probabilistic barrier remains the same.
What does "functional and selectable" mean? What would functional but not selectable mean, or selectable but not functional? Or selectable but not selected? Selection means no more, nor less, than the phenomenon by which a DNA sequence confers reproductive advantage on the phenotype. If it does so, it is functional, clearly. If it confers no net reproductive advantage, but does have an effect, you wouldn't normally call it a "function" in evolutionary terms. Now, clearly, when we observe microevolution (increase in finch beak size in response to changes in available seed size, for instance) we are seeing functional changes - an increased prevalence of alleles that tend to promote larger beaks. Over time, if El Nino events become more common, any new alleles that tend to promote even larger beaks will also become more prevalent. Mean beak size will continue to increase until still larger beaks are no longer advantageous. At what stage would you determine that 150 bits of functional information had been added? And how would you quantify the increase in each generation? Recall that the mean beak depth steadily increases from year to year.
Elizabeth: Microevolution is a random variation in the range of what a biological system can achieve, that gives a functional selectable result. Some forms of antibiotic resistance are microevolution, and they are well documented. A single aminoacid change can confer antibiotic resistance and be selected, in the presence of the antibiotic. That is well known, and observed. One aminoacis is about 4.32 bits of information. That is in the range of routine variation in a bacterial culture. 150 bits corresponds to about 35 AAs. A transition requiring 35 AAs to confer a new function has never been onserved to occur in any biological context.
So what makes you think that any new function requires a single step of 35 new AAs or AA changes?Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
08:05 AM
8
08
05
AM
PDT}

Let’s say, more correctly, that the trensition has a dFSCI of 150 bits. A1 would be the existing protein that “evolves” to A2. A2 would be a new protein, with a new function, a naturally selectable new function.
So let's get this absolutely clear: A1 is a DNA sequence that specifies a protein with a certain function, without which the phenotype would be worse off, right? A2 is a variant of that sequence that specifies a variant of that protein that serves a different function, without which the phenotype would be worse off? And A2 has 150 more bits than A1? Please correct me if I have this wrong.
A2 would differ from A1 of 150 bits of functional information.
OK. The best way to imagine that is that A2 differs from a! in 35 AAs that must necessarily change exactly to the new value for the function to emerge.</blockquote So you are proposing that sequence A1 is vital, and that A2 is an even better replacement for A1, but has 35 different AAs, that must all be present for the protein to be made at all?
As this is not usually the case, we can apply a Durston style computation, attributing to each site that changes an informational value in Fits that corresponds to the reduction in uncertainty that each AA site implies.
What is "not usually the case"? Are you suggesting that one DNA protein coding sequence mutates to a very different sequence coding for a very different protein in a single step? What evidence to you have for this?
So, at the two extremes, if one site must necessarily have one aminoacid, its Fit value will be 4.32 The function implies a complete reduction of uncertainty at that site. If instead any AA can stay at that site, its Fit value is 0 (no reduction of uncertainty is implied by the function). And similarly for all intermediate possibilities. The sum of all Fit values at the changing sites gives the dFSCI of the transition. When the starting protein (A1) is totally unrelated, as is the case for new basic protein domains, the total dFSCI of the new proteon can be approximated by the Durston method applied to its protein family.
Are you envisaging 35 new triplets or 35 changed triplets?Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
07:46 AM
7
07
46
AM
PDT}

I'm not "fight[ing] about words". I just wish that people would either use them more precisely, or say more precisely what they mean by them. Joe has been challenging evolutionists to explain "new body plans" for quite a while, and most of the responses have been in terms of hox genes. If, instead, he was talking about "new cell types" it would have saved a lot of angst if he'd explained that. However, now he has. Good.Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
07:32 AM
7
07
32
AM
PDT}

Not really a "personal dream" but clearly, if evolutionary theory is correct, they must have existed. So if you want to use them to falsify evolutionary theory you need to show that they did not exist. What you cannot do is say: look eukaryotes leapt into existence fully formed, therefore ID, unless you can show that that indeed happened. Nor can you say: unless evolutionist can show simpler earlier protoeukaryotes, ID is true. You can, however say: evolutionist don't have a good explanation for how complex eukaryotes emerged from simpler ones. And most evolutionists would be perfectly happy to agree. I'll repeat what I said to you in my recent post: evolutionists do not claim that there is no ID; they/we merely claim that there is no basis for an ID inference, and plenty of basis for our theory that complex functional living things can evolve by means of replication with heritable variation in reproductive success.Elizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
07:30 AM
7
07
30
AM
PDT}

Elizabeth:
No. There is your first mistake. There no warrant for the assumption that a new domain has to be advantageous (result in greater reproductive success for its phenotype than for phenotypes lacking the new domain). As long as it does not seriously impair reproductive success there is no reason why it should not “emerge” i.e. appear in one indivudual and be replicated down that lineage. There is no mistake. It’s simply you that don’t understand the reasoning. I am evaluating the probabilities of a certain final event to happen in a random system. The final event is supposed to be a naturally selectable function, because that would be the event that “stops” the random system and implies a nevessity mechanism (the expansion of the selected trait). Unless and until we get to that result, the system is random. And therefore we can evaluate the probabilities of a subset of states versus all possible states.
My position is that your reasoning is faulty. I also think it is confused. Firstly you assume a “final event”. There are no “final events” in evolution. It is a continuous process. And you don’t define “random system”. It’s not in itself a precise term. Do you mean “undirected”? Or do you mean “stochastic”? I will assume the latter for now, and parse your first sentence as: “I am evaluating the probability of a given event happening in a stochastic system.” And what do you mean by “naturally selectable function”? A DNA sequence that may at some point confer reproductive advantage on the phenotype? If not what? And why would such sequence “stop” the “random system”? And what do you mean by “the expansion of the selected trait”? Unless you can explain these things I cannot agree that “we can evaluate the probability of a subset of states versus all possible states”.
You have made that mistake many times. You seem to believe that neutral or quasi neutral variation can just happen, and that’s enough for you. Nothing could be fartgher from truth. It is obvious that neutral variation can happen,
Indeed it is.
but the problem is, it is random variation,
Again, what do you mean by “random”? Please give me a precise definition, because it is a word with a great many imprecise meanings and very few precise ones.
All states have the same probability to be reached by neutral variation.
No, they do not. Not all neutral variations are equally probable, and some states (again I’m not sure what you mean by “states” – the state of the DNA sequence? The state of the phenotype? The state of the population?) are therefore less probable than others. Also, some states may have no neutral-only pathways to them. And in any case, “neutral” is not an absolute term – a variant may be neutral (confer no reproductive advantage relative to phenotypes lacking that variant) in one context, advantageous in another, disadvantageous in another.
So, neutral variation is a perfectly random system where all unrelated states have the same probability to be reached.
So even without a clear definition of “a perfectly random system” this statement must be false. But it would certainly help if I knew what you meant by “a perfectly random system”.
But, if you want to defend an algorithm such as neodarwinism, that assumes that NS has an important role to explain how unlikely results are obtained in a random system, then you need to reach naturally selectable results. It’s as simple as this.
I’m not aware that “neodarwinism” is an algorithm, so I’m not attempting to “defend” it as such. But yes, of course, natural selection has an important role in evolutionary change – it’s the name we give to the process by which allele prevalence in one generation is biased in favour of those that confer reproductive success in the previous one. It is uncontroversial even amount IDists, who term it “microevolution”. It doesn’t explain “unlikely results” – it explains how traits that confer reproductive advantage in a given environment are likely to become more prevalent in that population – it explains adaptation, in other words. It’s not that “you need to reach naturally selectable results”. It’s that the resulting allele prevalence will be influenced by what confers reproductive success.
Nope. You just need the subset of sequences that result in a foldable protein that is not disastrously disadvantageous to the phenotype in the current environment. Same mistake as above, and even worse. What you say has no meaning. Why do you need a foldable protein at all?
I haven’t made a mistake, see above. And I was simply taking your case – how do you get a new protein domain? A protein domain is just a DNA sequence that codes for an amino acid chain with a 3D structure. So any DNA sequence that codes for such a thing for the first time will be a “new protein domain”, whether or not that sequence proves advantageous for the phenotype in which it first occurs. And as long as it doesn’t prove disastrous, it will be replicated many times in that lineage.
If neutral evolution takes place in a duplicated, inactivated gene, any sequence is neutral.
You seem to be very confused about the meaning of “neutral”. “Neutral” in genetics simply means “does not confer reproductive advantage on the phenotype in the current environment”. What that advantage is also relative – sometimes it is compared with the parental sequence, at its first appearance, sometimes with the sequence’s peers. The first might be neutral, the second advantageous, for the same allele.
You don’t even need the sequence to be an ORF. Any random sequence is certainly not disastrously disadvantageous to the phenotype, if it is not translated.
Exactly. So there is plenty of scope for new sequences to be replicated many times, even if they confer no advantage to the phenotype at first appearance.
If, on the other hand, the variation applies to a function sequence, any sequence that reduces or looses the original function will be visible to negative selection. See also my post 11.1.
Metaphors like “visible to negative selection” are very misleading. Negative selection is a way of describing a consequence, it isn’t an agent, and nothing is “visible to” it. It can’t “see”. If a sequence already confers some reproductive benefit to the phenotype then, clearly, a variation that disables that function will tend to result in that lineage (if we are talking about unicellular life) decreasing in prevalence. This means that sequences that confer reproductive benefit are more likely to be conserved. This is non-controversial, indeed, an essential strand in evolutionary theory.
Now you are double-counting, by equivocating with the word “function”, by conflating genotype with phenotype. If a protein is made, but has no effect on the reproductive success of the phenotype it doesn’t have a “function” in any relevant sense. It merely exists. And provided the first few generations down that lineage survive, it will exist in sufficiently many copies that it is likely to hang around for a long time. Always the same error. I am discussing the probabilities of reaching a naturally selectable result in a random system of variation, If you cannot understand that, you cannot understand any part of my reasoning.
It’s not that I don’t understand it, it’s that in my view, the error is yours! Or, at any rate, the confusion. Your position seems, as I suggested, to boil down to the IC argument – that some advantageous sequences can only be reached by a series of non-advantageous changes. And you are providing no evidence that a protein domain is IC.
Better to call it P(sequence-that-results-in-a-folded-protein), which may be a tiny subset of all possible sequences, but still may be quite high given the immense number of opportunities for sequence mutations to occur. No. Wrong! A folded protein is not functional and is not naturally selectable.
How do you know? There is no way of knowing, without observing the phenotype in which it occurs in the environment in which that phenotype lives. And, as I said, it doesn’t matter anyway, although if the folded protein in question confers some reproductive advantage, then it is more likely to be conserved. I think you need to unpack this term “naturally selectable”. I am not sure what you mean by it, and I suspect that what you think you mean by it will make no sense when you unpack it.
Until a functional selectable result occurs, a foldable protein is not different than a non foldable sequence that is not translated. And why should a foldable non functional protein be translated? If that were the case, all living cells would be repleted of foldable non functional proteins that are not “disastrously disadvantageous to the phenotype”. I suppose that’s not the case.
Clearly any sequence that does not result in any effects on the reproductive success of the phenotype is not selected ie. is neutral. By definition. That includes untranslated protein sequences, translated protein sequences that are not expressed in tissues in which they provide any selective advantage, and sequences that don’t do anything at all. As to “why should a foldable non functional protein be translated?” – the proximal answer is biochemical (if the required translational sequences are present and activated by other chemical signals, and the distal answer (the teleonomic answer) could be because phenotypes bearing sequences in which it is translated under the conditions in which it is translated reproduce better than those that don’t, resulting in greater prevalence of their genotypes in the population.
And we simply do not know how large that subset of sequences is, nor indeed whether some of their precursors also result in reproductive advantage for their bearers. That’s really an “argument from ignorance”, if I ever so one! Yes, we don’t know, because nobody has ever been able to show those precursors, either in the proteome or in the lab. But you can always hope and dream…
Gpuccio, it may well be an “argument from ignorance” but the argument is simply: you cannot compute the probability of an event unless you know what the probability space is. If you are ignorant of the probability space, then you can’t compute it. Therefore you can’t infer design from the low level of that probability. It’s not evolutionists who are making probability arguments, it’s IDists. And you can’t compute a probability in ignorance! Not without making a Bayesian stab at the likelihood that your priors are correct.
It isn’t precise at all. It’s wrong in a number of respects (the sequences coding for a new domain don’t have to be currently advantageous to appear, and there is no good reason to assume the precursors of those sequences don’t confer reproductive advantage), and we don’t in any case know the size of the subset of DNA sequences that result in foldable proteins, though we may know roughly the size of the subset that have actually appeared in living things. You are simply misrepresenting my argument. I never said that a sequence has to be “currently advantageous” to appear. That’s only your imagination.
OK, cool. I wasn’t intentionally misrepresenting your argument, I’m just trying to make sense of it. It still makes little sense to me for reasons I have given above.
What I said is that a sequence has to be currently advantageous to be naturally selected, and that up to that point all sequences have the same probability to appear. Is that the same thing, in your opinion?
Well, clearly a sequence has to be currently [reproductively] advantageous to be naturally selected because those two things are synonymous. Your second clause is simply wrong. Not all sequences have the same probability of appearing. I’m not even sure what you mean by that, but I can’t think of any sense in which it could be true. I think one of the problems with your approach is that you don’t specify your priors. Given, for example, the (non-functional) sequence: ABCBBCDADA which is more probable as the next variant: ABCBBBDADA, ABCBBCDADADA, or BCDDADCBAB?
The size of the subset of foldable proteins is interesting, but not relevant. The only relevant subset is the one I defined, the subset of naturally selectable proteins. All the rest is random variation.
Well, you need to define “naturally selectable” proteins, and also how you identify the subset.
This exactly the problem Petrushka mentioned, and which IDists tend to dismiss as ignorance because they think they’ve dealt with it. You haven’t. We have dealt with it, and we will continue to deal with it. Because it is an importnant problem, and we believe it has to be dealt with. That does not mean that the problem is completely solved, obviously.
It’s neither solved nor solvable. Therefore not only does it not yield the claimed Design inference, it’s a fruitless approach to detecting design.
Darwininsts, if they were scientifically honest, should deal with it with the same urgency (some, indeed, have tried, with terrible methodology and false results). Because the problem is fundamental for their own theory.
It has nothing to do with scientific honesty and everything to do with a sound understanding of probability and the scientific method. You cannot falsify a scientific hypothesis by concluding that the observations are “improbable” given the hypothesis. You can only falsify a null hypothesis that way. Evolutionary theory is not a null hypothesis.
Instead, most darwinists, including you, just try to hide behind the supposed impossibility to solve the problem, so that their wrong theories may conrinue to be believed for some more time.
This is not true. It’s the whole approach that is misguided. There are plenty of ways of doing research into evolution that do not involve trying to compute the probability that it did not happen that way, which is impossible. To argue that we are “hid[ing] behind the supposed impossibility” is ludicrous. It’s just not how science is done. It doesn’t work. It’s invalid.
The problem can be solved, and we have a lot of indications about what the solution is. And the solution is exactly what ID has shown.
I have tried to demonstrate to you why a) the problem is not a problem, b) it can’t be solved and c) it wouldn’t tell you anything even if it could be. The problem is ill-posed. In fact, to be really frank, it isn’t posed at all. If you disagree, please pose it, precisely :)
You have pre-defined the target as those DNA sequences that result in foldable proteins that form part of modern functional proteins, and assume that that small target comprises the only possible target, forgetting that there may be a vast set of DNA sequences that would also result in protein domans, and another vast set of proteins comprised of those domains that, in some alternative universe, might also prove to confer reproductive advantage in some alternative biosphere. Wow! Have you lost your mind? I have done nothing like that. I have defined a subset as (I quote myself): “That means that we have to look for a specific subset of P(F) (the subset of folded proteins), the subset of folded proteins unrelated at sequence level to already existing proteins in the proteome, with a new fold and a specific new biochemical function. Let’s call that P(NUFF), for New Unrelated Folded and Functional. Then we have to look for an even smaller subset of that, the NUFF that are naturally selectable in that context, IOWs that can confer, by themselves, a reproductive advantage in the context of the living being where the transition is supposed to happen (prokaryotes, I suppose, or a symbiosis of them). Let’s call that P(NUFFNS).” Where in that is the concept you attribute to me? Just to be precise, “those DNA sequences that result in foldable proteins that form part of modern functional proteins? Why do you put in my mouth things I have not said?
As I said, I’m trying to understand you. Your words, as written, make no sense to me. I don’t know what you mean by the terms you are using (such as “naturally selectable” and “random system” ) so I’m having to make tentative assumptions. Clearly erroneous ones, but I hope at least it is becoming clear why you are unclear to me ?
I will not comment about the “alternative universe” and “alternative biosphere” part, just out of respect for you. I believe you must be really desperate to use those arguments. And anyway, what in the world does an alternative universe have to do with the probabilities of a selectable function arising in a specific biological context, such as my example of prokaryotes transitioning to eukatyotes?
No, I am not “really desperate”. Boy, this has come round full circle. I’m simply pointing out that you haven’t come close (and cannot ever come close) to defining the probability space. That’s why your argument suffers from the exact flaw Petrushka pointed out: you restrict the “target” space to something only a little larger than what is observed, and then express astonishment at the low probability that this tiny space should have been hit. You have absolutely no basis on which to define that target space so narrowly – indeed it is uncomputeable, as I have said. I wasn’t talking about multiverses or anything abstruse like that, just what might have happened if what actually happened didn’t. After all, your own existence, you, gpuccio, in all your uniqueness and complexity, might never have existed, were it not for a whole series of events that could easily have taken a different turn. How do you know that a completely different set of protein domains were just as viable as the ones we observe? Or countless sets of protein domains?
And this, to repeat, is the fundamental error at the heart of ID: to look at what exists, and say: what exists must be a tiny subset of what might exist (correct), and is also the only subset possible that could result in life (incorrect) and is therefore extremely improbable. Which is fallacious, because based on a false premise. Wow again! Where have I, (or ID, for what I know), ever said that “what exists is the only subset possible that could result in life”? I hyave never said that, I am really sure of that. For the simple reason that I don’t believe it, and I usually don’t say things I don’t believe.
Good. Then how do you compute the subset?
But I have sais a lot of times thatalready existing information and complexity poses huge constraint to what can be useful in that context. IOWs, if you have to find something useful in an existing strain of bacteria, your options are radically limited, and the possibilities that other forms of life based on fire and lithium could possibly exist in another galaxy are scarcely a help!
Yes, indeed (and I did not have in mind “life based on fire and lithium” but only, in this context, life based on a different set of protein domains), but you have it backwards. Already existing information and complexity does indeed pose a huge constraint” but not just on “what can be useful in that context” but on what variants will be viable in that context. For example, short arms may be advantageous relative to no arms, but once short arms exist, only longer arms may offer any advantage, and once longer arms exist, short arms may be disadvantageous. And, at a molecular level, once a sequence confers a reproductive advantage it will be highly conserved, meaning that only variants that confer increased reproductive advantage will tend to be retained in the population. “Natural selection” therefore is a kind of ratchet mechanism by means of which incremental improvements in adaptation to an environment can be steadily gained. You are correct, of course, in saying that in bacteria, this imposes particular constraints, because horizontal gene transfer mechanisms are limited and haphazard. However, once sexual reproduction got going, whole new avenues opened up, because now sequences could be mixed and matched, genes could propagate independently of the rest of the genome, giving rise to much more effective filtering of advantageous sequences.
Quite apart from the fallacy that only sequences that confer reproductive advantage on the phenotype can be replicated, Totally invented fallacy. I never said that.
Good. In that case I’m not sure what you meant when you said that you needed to “reach” a “naturally selectable” sequence. I hope you can explain that. I’m doing my best here, gpuccio, and I apologise for my part in misunderstanding you, but I do think that the major part of the misunderstanding arises from confusion in your own thinking.
and the fallacy that a sequence that confers a new advantage can have had no precursors that also conferred an advantage (in some different way). Totally invented fallacy. I certainly said, and say, that those precursors have never been shown. And I have said many times that science is note made on “mere possibilities”. It usually needs some facts, you know?
You cannot make an inference from lack of evidence. If there can be advantageous precursors, then the thing is not “irreducibly complex”. If you want to infer design from “irreducible complexity” then you need to show that there can have been no advantageous precursors. The onus is on the one making the claim. Evolutionist aren’t claiming that there was no intelligent designer. It is IDists who are claiming that there was.
And the general confusion between genotype and phenotype. What confusion? I have no confusion about that. That is a serious accusation. Please, detail it.
It seemed to me apparent in your use of terms like “function” and “naturally selectable” and “sequence”. If you are not confused about that, then perhaps the confusion arises from some other source. But you certainly, in my view, need to sort out what you mean by those two terms before going any further. Anyway, nice to talk to you (if only to disagree with you!) again! A happy new year to you LizzieElizabeth Liddle_{December 30, 2011
December
12
Dec
30
30
2011
07:21 AM
7
07
21
AM
PDT}

gpuccio,
"That is the tiny subset of probability we are looking for if we want to believe in a darwinian explanation of the emergence of a single new basci protein domain in the course of natural history (an event that took place at least 2000 times in natural history)."
That's how it looks to me as well. The probabilities of course multiply as each of those protein domains are considered. Even if it were granted that NUFFNS was close in size to F (the folding set), it's difficult to see how a random search could yield much success.material.infantacy_{December 30, 2011
December
12
Dec
30
30
2011
07:19 AM
7
07
19
AM
PDT}

It's not only an offence to Petrushka's own intelligence - it's a science stopper. To any question "Why is this fact so?" you don't look for explanations that are more or less likely. You just say, "It's a fact, so it had to happen - probability = 1" You find the Lord's prayer inscribed on Precambrian sediment? Any random pattern is equally unlikely in advance. But this one has already happened, so it has a probability of 1. Mathematics, or sophistry?Jon Garvey_{December 30, 2011
December
12
Dec
30
30
2011
07:12 AM
7
07
12
AM
PDT}

Elizabeth: I don't want to interfere, but there is probably no reason to fight about words, while it is possible to agree on a shared meaning. It is true that "body plan" is usually referred to metazoa, but it is also true that "cell type" is also used to distinguish between different cell types in a multicellular being. It is certainly true that, for unicellular beings, that cell is the body. And it is true that it has an inner organization. That we call it "body plan" or not is a free, and not so relevant, choice. That the inner structures of prokaryoyes and of unicellular eukaryotes are very different is, I believe, a fact.gpuccio_{December 30, 2011
December
12
Dec
30
30
2011
06:38 AM
6
06
38
AM
PDT}

Prev 1 … 3 4 5 6 7 8 Next

You must be logged in to post a comment.

Leave a Reply