How Random is Random Mutation?

Hawks: “Of course copying can produce “new, readable code”. In what sense is agata + ct = agctata (as an example of the insertion of ct into agata) not the production of “new, readable code”. agctata was not there before, therefore the code is new. It is, also, still readable.” If that insertion was in a coding gene then you just broke it because you need to insert whole codons (triplets). Your insertion shifted the reading frame and turned most subsequent amino acid specifications into a different acid. Of course you knew that, right? :roll: DaveScot

I've made it back to this post while looking for something I had written previously about transposons. I think I need to respond to a couple of posts above. First, HodorH writes: "Hmm, you seem to be doing calculations for a constant population size of 1. My above objection obviously applies to this section as well. " This is in response to a calculation I had made to determine how many generations would be needed to develop two, side-by-side, missense mutations. Now, as I admit on the post, I was under the impression the authors meant a two nucleotide change when they were referring to a "missense" mutation whereas the authors were dealing with SNPs that led to the "missense" mutations. But my calculation is valid for two side-by-side mutations even in yeast populations of great size. The reason for this is that it is unimportant how quickly "one" mutation occurs at the necessary site in ONE yeast cell--that can, and does, occur quite rapidly because of population size and the given mutation rate of DNA; however, once a "single" mutation has occurred at a site in ONE of the cells, THAT PARTICULAR CELL will be the fastest, statistically, to generate the second needed mutation adjacent to the first. Hence, we only need to follow it's progeny. So, yes, it was only ONE cell for which the calculation was made. Hawks: "Of course copying can produce “new, readable code”. In what sense is agata + ct = agctata (as an example of the insertion of ct into agata) not the production of “new, readable code”. agctata was not there before, therefore the code is new. It is, also, still readable." It appears you don't understand the meaning of the word "copy". "Copy" doesn't mean "insertion". "Doesn’t matter. An insertion, such as illustrated above, no matter how it is done still creates “new, readable code”." But, indeed, it does matter. Indeed, the very fact that an insertion can be "read" at all strongly suggests that the insertion is "directed", since if this were ordinary binary code, placing an added 00, or 01, or 10, or 11 into a string of code could severely affect the readability of the code. Your later comment about the negative effects of mutations suggests that those who survive, survive, and those which don't, don't, and, of course, its the surivival fo the survivors that were interested in, isn't it. Well, we can be non-chalant about it when it comes to yeast--they live in huge populations and mulitply instantly. Yet this same phenomena happens in humans, and human populations would never survive the "survival of the fittest" game if these transposon movements proved to "randomly" be "readable or not". trrll: "If, almost every time, you find him at the top of the nearest hill, would you conclude that he has been “directed?” " Regarding your thought experiment: that the person is not found in a valley IS "directed"--because that's, in essence, what you directed him to do. But the fact that over time he may be on ANY ONE of the hilltops shows that the process is "undirected". So the moral of the story seems to be: if you provide information, results follow; if you don't provide information, then randomness ensues. ofro: "I don’t know what that is supposed to mean. The only information that I can see in an atom or molecule in its ground state is the nuclear and electronic configuration that identifies it as that molecule or atom. Or are you referring to the highly artificial conditions of quantum computing? Then you are way off. Or DNA computing? Way off again. All this is pure physics and has nothing to do with any biological mechanisms found in cells. Don’t forget: this is a biological system which imposes pretty strong restrictions on the theoretical physical effects that physicists and engineers may be thinking of." All things that exist, exist in a certain quantum state. In fact, you mention the ground state. But each individual atom can become excited away from its ground. And when atoms interact with other atoms, adjustments are made to their ground states. Now if you study quantum mechanics, you soon realize that the only "system" that QM can deal with mathematically, more precisely, analytically, is the hydorgen atom: one electron, one proton--as simple as it gets. Hence, ONE atom has more quantum information contained within it then any scientist can precisely determine. That is why quantum computing needs only a small number of atoms to do huge computational jobs. But quantum computing is itself a way of acknowledging the informational complexity that makes up even simple compounds. As to whether this has anything to do with biology, it's fascinating that Schrodinger, giving his Nobel Prize Address, stimulated James Watson (it could have been Crick--can't remember for sure) in his search for DNA since Schrodinger, in "What is Life?" (his address) suggested a mechanism by which information could be stored and changed biologically--and the explanation was at the level of cosmic rays interfering with some crystalline structure--a change brough about via quantum mechanical effects. The force of my argument is that "information" (quantum mechanical information, if you will) subsists in all of reality and is its engine-----but, of course, all of this is looking into the future. Related to this, here is what you, ofro, wrote elsewhere: "The problem I have with any telic mechanism, as I pointed out earlier, is the lack of a mechanism that could direct a mutation. The information that is needed to “know” the consequence of each possible mutation is simply way too large to be stored away somewhere in the genome. " Well, if 30 atoms, entangled quantum mechanically, can outperform any computer man has made to date, then I wouldn't worry about how much information is neede. PaV

PaV wrote: "In the sense that ‘copying’ does not produce new, readable code. It’s simply copying it. " Of course copying can produce "new, readable code". In what sense is agata + ct = agctata (as an example of the insertion of ct into agata) not the production of "new, readable code". agctata was not there before, therefore the code is new. It is, also, still readable. "Added to this is the fact that we don’t know what mechanisms are involved in the shuffling around of DNA sequences." Doesn't matter. An insertion, such as illustrated above, no matter how it is done still creates "new, readable code". "Is it random? It sure doesn’t make sense that it would be." It doesn't make sense that it would HAVE to be. " Would you want sections of your software code to move around at random in your computer? I think not. " What you seem to be alluding to here is that moving stuff around is more likely to be "bad" rather than "good". Given the experimental evidence that suggests that most mutations (that actually affect fitness) are negative, this would also seem to be what happens in biological organisms. It does not matter one iota if I want it or not. Hawks

Corrections: In the last post, in quoting Orr, it should read "opening sentence", not "opening paragraph." And it should read, "the reasons a mature theory....", and not "the reasons a mutre theory....". As well, there's a certain section in blockquotes (the middle of the triple blockquote section) that should not have been: it represented my response. Sorry for the errors. I'm watching the U.S. Open. PaV

I don't have the time, or the inclination, to be going round and round on all of this. But, let me respond: trrll: "Not gene duplication and deletion, but gene duplication and then selection of one of the copies for a modified function." What evidence is there that this 'modification' has been brought about by random processes? Behe and Snoke make it readily apparent that this kind of change is impossible via random mutations. Just finding that a gene is modified, doesn't tell us 'how'; and, of course, it is the 'how' that is the most important thing to know.

Sometimes, when I write a paper, I copy a previous paper that I’ve written, keep the parts that apply to the new topic (perhaps historical overview of the field as a whole), and revise the rest appropriately.

What you copy is not new information. What you revise is the work of an intelligent agent, hence, new information.

Originally, a bit of random sequence that happens to have a useful function.

If it is a random sequence, then how could it have a useful function. We're dealing with coded language. If you put a random string of 0's and 1's into a computer, you'd have no useful function whatever.

Biologists have been discussing these issues in research papers for decades. It is indeed part of modern evolutionary biology as biologists understand it, even if you are a bit behind the times.

Am I behind the times? Isn't it evolutionary biologists who are behind the times? Where is the theory? I'm familiar enough with the "discussions"; where is the theory? Or is a theory just plain impossible? Here's the opening paragraph of a paper by Allen Orr in Nature last year: "Adaptation is not natural selection." And not too later: "Here I survey these attempts [to answer some important questions about the genetic basis of adaptation]. My approach is historical, consdiering the rise and fall of various views on the genetic basis of adaptation, the reasons a mutre theory has been slow to develop and the prospects and problems facing current theory." In his conclusion, he writes: "There are in fact two problems. The first is that current theory is limited in several ways--all the models that have been mentioned rest on important assumptions and idealizations. Fisher himself noted such limitations in correspondence about his geometric model. And the mutational landscape model assumes that adaptaion occurs at a single gene (or small genome), that fitness distributions show certain tail behaviour, and that mutant fitnesses are taken fromt he same distribution throughout adaptation. Although they are reasonable starting points for theory, none of these assumptions is necessarily correct and changing any might well change our predictions." Again, where is the theory?

But we are seeing examples of entirely new activities evolving, such as nylonase.

You're talking about bacteria, not yeast. And, please tell me how you know that this capacity 'evolved'. We know that certainly there are bacteria that 'digest' nylon, and the argument goes that since nylon wasn't invented until 1935, how could the 'gene' be present. Nylon is a polyamide. There are certain chemical bonds present. We know that bacteria can break down almost anything in nature. How can we distinguish between a gene for breaking down the polyamide bonds being present within the genome ab initio,and the 'gene' developing de novo? Is every changed situation in an organism 'evolution' by definition?

“Wander around randomly, but you are only allowed to walk uphill.

This is part of your 'thought experiment'. Excuse me, but isn't telling someone they can't go uphill giving them directions as to what they can, or cannot do? In other words, "all roads to galactose metabolism lead to Gal80."

Or to put it another way, the presence of a selection rule (e.g. “walk uphill”) gives a random process direction.

That's one way of looking at it. But I'm suggesting that having only one way to metabolize galactose (i.e., the activation of the Gal80 gene) allows an organism to employ random mechanisms in adapting to varied environments. Hawks:

Again, what definition of information are you using?

In this thread I've used two definitions: one at a quantum mechanical level, and one at the level of code language (DNA). In the quote you reference my definition of information is readable code.

You seem to be implying that new information has to be encoded within a protein-coding gene.

I'm saying that for a gene to come into existence, information (readable code) is necessary.

In what way is this NOT information generation?

In the sense that 'copying' does not produce new, readable code. It's simply copying it. Added to this is the fact that we don't know what mechanisms are involved in the shuffling around of DNA sequences. Is it random? It sure doesn't make sense that it would be. Would you want sections of your software code to move around at random in your computer? I think not. PaV

PaV wrote: "In the sense that they don’t create GENES! Look at the mechanisms you’re proposing: (a) gene duplication and deletions. How does deleting a gene create information? How does duplicating a gene create information? It copies it just like I can copy a research paper on my copier, but that doesn’t create any new information that wasn’t in the original research paper to begin with. (b)transposons, and insertion sequence: In both cases you’re dealing with sequences that already exist and that are being moved around; (c)integrons/gene casettes: Gene cassettes are DNA sequences that encode one or more genes resulting in a single biochemical function. But again, the ‘genes’ are already there. In sum, whence the ‘genes’. That’s the question Bateson asked at the turn of the 20th Century. Any answers? If one wants to invoke point mutations, then one runs right into the arguments of improbability that ID presents, and the results of Behe and Snoke’s paper." Again, what definition of information are you using? I sure am not familiar with the one you are using. You seem to be implying that new information has to be encoded within a protein-coding gene. This is certainly not an accepted definition. Even if this is a useful definition, it does not in any way change the fact that all these mechanisms can create information. Gene duplications and deletions don't by necessity delete/copy entire gene - merely parts of them. The word gene duplication/deletion should more correctly be called DNA duplication/deletion. Likewise, the other mechanisms are able to insert into other genes, creating novel such. In what way is this NOT information generation? Hawks

In the sense that they don’t create GENES! Look at the mechanisms you’re proposing: (a) gene duplication and deletions. How does deleting a gene create information?

Not gene duplication and deletion, but gene duplication and then selection of one of the copies for a modified function. I wouldn't say that this creates information, however; all of the information necessary to define every possible organism is inherent in the laws of nature. It is more accurate to say that selection transfers information from the environment to the genome. That this must be the case is clear, because in some cases, it is possible to determine information about the environment (e.g. what nutrients are present) by examining the genome.

It copies it just like I can copy a research paper on my copier, but that doesn’t create any new information that wasn’t in the original research paper to begin with.

Sometimes, when I write a paper, I copy a previous paper that I've written, keep the parts that apply to the new topic (perhaps historical overview of the field as a whole), and revise the rest appropriately.

integrons/gene casettes: Gene cassettes are DNA sequences that encode one or more genes resulting in a single biochemical function. But again, the ‘genes’ are already there. In sum, whence the ‘genes’.

Originally, a bit of random sequence that happens to have a useful function. That random sequences can exhibit useful catalytic activity has been demonstrated in numerous experiments evolving novel enzymatic activities from random protein or RNA sequences.

The point is this: there are genetic effects now known that, on the surface, seem to have more powerful effects than mere point mutations, and yet there has been no comprehensive reformulation of the Modern Synthesis. Why not?

You probably just haven't been paying attention to the literature. Biologists have been discussing these issues in research papers for decades. It is indeed part of modern evolutionary biology as biologists understand it, even if you are a bit behind the times.

Now, if we were dealing with a kind of yeast that had no gene for metabolizing galactose anywhere in the genome, and, through random mutations a gene to do so is somehow constructed, then, yes, we all take off our hat to RM. But that’s not what we’re seeing.

But we are seeing examples of entirely new activities evolving, such as nylonase. But remember that natural selection's natural mode of invention is tinkering, so if their is a "short cut" to solve a problem, such as making more or less of something that the cell already makes, there is a strong chance that evolution will take that path.

f it’s hardly surprising that a mutation in a repressor for the Gal80 gene is the most readily discovered solution to this problem, then this suggests that the ’solution’ is directed–after all, all four found the, more or less, ’same solution’.

Let's do the following thought experiment. Take somebody out into the country (let's pick an area of gently rolling meadows, to be safe). Blindfold him, spin him around a few times, and give him the following instructions. "Wander around randomly, but you are only allowed to walk uphill. When you get to a point when you can't find any direction that is uphill, stop." Repeat the experiment a few times, always starting at the same place. If, almost every time, you find him at the top of the nearest hill, would you conclude that he has been "directed?"

Yet, the activation mechanism seems to be ‘random’ point mutations. So, it seems as though ‘randomness’ is the active agency in bringing about this ‘directed’ solution. But, based on what I’ve written above, maybe the biggest surprise of all, and the most amazing, is not “that in a selective environment, random mutation produces results that superficially look as if they are directed,” but rather, that what is a ‘directed’ process looks to be ‘random’.

Or to put it another way, the presence of a selection rule (e.g. "walk uphill") gives a random process direction. trrll

Hawks wrote: In the sense that they don't create GENES! Look at the mechanisms you're proposing: (a) gene duplication and deletions. How does deleting a gene create information? How does duplicating a gene create information? It copies it just like I can copy a research paper on my copier, but that doesn't create any new information that wasn't in the original research paper to begin with. (b)transposons, and insertion sequence: In both cases you're dealing with sequences that already exist and that are being moved around; (c)integrons/gene casettes: Gene cassettes are DNA sequences that encode one or more genes resulting in a single biochemical function. But again, the 'genes' are already there. In sum, whence the 'genes'. That's the question Bateson asked at the turn of the 20th Century. Any answers? If one wants to invoke point mutations, then one runs right into the arguments of improbability that ID presents, and the results of Behe and Snoke's paper. Ofro: "When these researchers did their work, the phenomenological basis of inheritance (i.e. Mendel’s observations) was just-rediscovered. The hereditary material was referred to as germ plasm and nobody even knew yet what DNA was. The point is this: there are genetic effects now known that, on the surface, seem to have more powerful effects than mere point mutations, and yet there has been no comprehensive reformulation of the Modern Synthesis. Why not? Ofro: "Don’t distract by talking about “information content”. I was comparing the effect of different kinds of random mutations on adaptation to a changed environment." Why do you consider this a distraction? Isn't it the most critical question of all? Yes, you're invoking random mutations as a mechanism of adaptation. But the effect of random mutation is to activate a mechanism that already exists. Now, if we were dealing with a kind of yeast that had no gene for metabolizing galactose anywhere in the genome, and, through random mutations a gene to do so is somehow constructed, then, yes, we all take off our hat to RM. But that's not what we're seeing. In some of what I later wrote I was suggesting that the yeast genome may be so constructed that its activation regions are designed to switch from 'off' to 'on' through the agency of RM. Just because a random effect is occurring, that doesn't mean that what ensues is randomly occurring. Think of Schrodinger's Cat: a random effect (Beta decay) triggers the release of poison from the vial inside the box, thus killing the cat. Killing the cat is purposeful, though a random effect is part of the overall process. Ofro: "Don’t distract by talking about “information content”. I was comparing the effect of different kinds of random mutations on adaptation to a changed environment." There was a paper that came out earlier this year that discusses this phenomena. Sorry I don't have a reference for you. I took a quick look through Google, but no luck. Maybe somebody else remembers that paper. I think it was discussed on this board, at least briefly. "Are you insinuating that a mutation in the DNA is changing its information content? I would be quite happy to hear that since that also means that certain mutations can increase the information content of the genome." Yes, I am. However, this 'information content' is very small. We're basically talking about an "on/off" switch. In binary code, that's simply either "0" or "1". So the 'bit size' of this information content is 1 bit. Tiny. The 'bit size'(information content) of a DNA region coding for a protein is, on the other hand, huge; hence, point mutations are of no use, probabilistically, in producing these sequences randomly. You can read the ID literature to see why. Ofro: "The only kind of information in DNA you can convince any knowledgeable scientist of is the primary nucleotide sequence, with possible additional effects by chemical modification such as methylation and secondary structural effects such as supercoiling. If I left anything out, I’ll be happy to hear about it and potentially stand corrected." If you mean that all information is contained in genes, then we only have to consider the fact that Chimpanzees and homo sapiens are 97% the same when it comes to the genes found in our genomes. Here's what trrll wrote in a post right after yours: "Indeed, a lot of evolution is clearly at the level of regulation. A chimp has basically the same genes that we do, and they do the same things. But they are regulated very differently." Now, if what you mean is that the primary nucleotides code information, then isn't it straightforwardly obvious that if a mutation occurs, then the information content has occurred. The question, then, is only about the degree of that information change, which I addressed above. Ofro: "And just think how much DNA sequence would be needed to account for all the possible contingencies. " In a sense, the whole point of what I wrote was to address this very point. What I'm suggesting is that any one point mutation will not change the overall energy content (using the term 'energy content' as a way of describing changed quantum mechanical states) of the transcribed regulatory region of particular genes. However, it is possible that these regulatory regions can be so constructed that a point mutation in a particular location of that region (a "critical" location) might, indeed, effect a "critical" overall change in its energy content, thus 'triggering' the regulatory region to become activated. In other words, if I'm a designer, and I know that 'random mutations' can occur, I can so construct a region so that, given the right conditions, these 'random' effects can be utilized so that an organism can adapt to changed conditions. This, I know, is speculative. It is also somewhat predictive. In other words, if, indeed, the genome is 'designed', then eventually it should be demonstrated that this kind of 'control mechanism' actually involves 'random mutations' as part of its adaptive function. You see, ID really does make predictions--unlike Darwinism, where everything is explained ex post fact. trrll: "Genes aren’t just moved around, they are sometimes duplicated, effectively creating a new gene. Rearrangements can also move a part of one gene into another, creating a new gene by gene fusion." But, indeed, genes do move around. In fact, they jump. In this case, from Chromsome #4 to Chromosome #3. Here’s the link: http://www.newscientist.com/article/dn10027-frankenstein-protein-defies-biology-textbooks.html trrll: "What this thread comes down to is amazement that in a selective environment, random mutation produces results that superficially look as if they are directed. . . . . . . . . It is hardly surprising that a mutation in a repressor is the most readily discovered solution to this problem." It seems like we have a stand-off. If it's hardly surprising that a mutation in a repressor for the Gal80 gene is the most readily discovered solution to this problem, then this suggests that the 'solution' is directed--after all, all four found the, more or less, 'same solution'. Hence, the only way to survive seems to be through Gal80 gene activation. There's directedness here. Yet, the activation mechanism seems to be 'random' point mutations. So, it seems as though 'randomness' is the active agency in bringing about this 'directed' solution. But, based on what I've written above, maybe the biggest surprise of all, and the most amazing, is not "that in a selective environment, random mutation produces results that superficially look as if they are directed," but rather, that what is a 'directed' process looks to be 'random'. PaV

PaV:

Chromosomal rearrangement moves genes, but doesn’t create them.

Genes aren't just moved around, they are sometimes duplicated, effectively creating a new gene. Rearrangements can also move a part of one gene into another, creating a new gene by gene fusion.

We’re not seeing a change in the galactose metabolizing gene–it remains unchanged; all that is changed is the ‘off-switch’, if you will.

Indeed, a lot of evolution is clearly at the level of regulation. A chimp has basically the same genes that we do, and they do the same things. But they are regulated very differently. What this thread comes down to is amazement that in a selective environment, random mutation produces results that superficially look as if they are directed. Of course, this is not a surprise to any biologist. The frequency with which a particular type of mutation arises in response to a particular pattern of selection will depend upon how many different ways there are to improve survival in that environment, and their relative probability (e.g. multi nucleotide changes having a lower probability than single nucleotide changes). If most of the pathways to improved fitness involve Gal80, that will be most common. It is hardly surprising that a mutation in a repressor is the most readily discovered solution to this problem. There are usually more ways to break something than to improve it, so if there is a way to solve a problem by disabling a specific protein, that will tend to turn up first. In this case, mutation/selection discovered 3 different mutations that disrupt the function of the repressor. trrll

(my apologies if my effort to print in italics is failing and messing up the print) ofro: “But I see no conceptual difference between a randomly occurring point mutation and a randomly occurring chromosomal rearrangement. PaV: “I couldn’t disagree with you more. The Modern Synthesis is based on the mathematics of Fisher, Haldane and Wright, none of which included the notion of chromosomal rearrangements and such.” I agree that non included the notion of chromosomal rearrangements; the reason is that they had never heard of it. According to Wikipedia, “Modern Synthesis” isn’t quite that modern, having been developed at the beginning of the 20th century. When these researchers did their work, the phenomenological basis of inheritance (i.e. Mendel’s observations) was just-rediscovered. The hereditary material was referred to as germ plasm and nobody even knew yet what DNA was. “So, Darwinism can invoke all the “random” mutations and rearrangements they want, but until such time as a plausible mechanism for information generation is presented, it will remain what it currently is: vacuous.” Don’t distract by talking about “information content”. I was comparing the effect of different kinds of random mutations on adaptation to a changed environment. “As I mentioned in my last post, mutations have a way of canceling out the effect produced by previous mutations.” What is the origin of this notion that mutations cancel other mutations? If you are referring to the revertants that I mentioned in post #18, these are just as rare as, if not rarer than, the original mutations. And for every revertant there are several mutations that enhance the original mutation, which, overall, makes correcting a mutations much more difficult than achieving the first mutation to begin with. On top of that, I don’t know to what your response is supposed to address. My comment was about the telic nature of the mutations that you are supposing in order to achieve adaptation, and you are referring to a mechanism that is supposed to counteract such a mutation. What then was the purpose of the first mutation? “Science is on the way of demonstrating in a clear fashion that each molecule, each atom, contains quantum information, and that molecules themselves can act as almost a computer, adjusting its information content based on the chemical/physical changes that occur to them.” I don’t know what that is supposed to mean. The only information that I can see in an atom or molecule in its ground state is the nuclear and electronic configuration that identifies it as that molecule or atom. Or are you referring to the highly artificial conditions of quantum computing? Then you are way off. Or DNA computing? Way off again. All this is pure physics and has nothing to do with any biological mechanisms found in cells. Don’t forget: this is a biological system which imposes pretty strong restrictions on the theoretical physical effects that physicists and engineers may be thinking of. “adjusting its information content based on the chemical/physical changes that occur to them.” Are you insinuating that a mutation in the DNA is changing its information content? I would be quite happy to hear that since that also means that certain mutations can increase the information content of the genome. “Within this kind of a scenario, it would be possible to program a response to a “critical” (and this is sort of the operative word here) change that might occur in “critical” molecules (notably proteins). You lost me. What kind of response programming are you referring to? Something that nature supposedly has programmed into the DNA or something that humans might do at some time in the future? The only kind of information in DNA you can convince any knowledgeable scientist of is the primary nucleotide sequence, with possible additional effects by chemical modification such as methylation and secondary structural effects such as supercoiling. If I left anything out, I’ll be happy to hear about it and potentially stand corrected. So if there is any information in the cell on how a mutation (not a metabolic regulatory pathway like that in response to a switch to galactose) might help in adapting to an environment it is with overwhelming probability bound to be found in the DNA sequence (of course, the metabolic pathway is encoded as well). And just think how much DNA sequence would be needed to account for all the possible contingencies. ofro

PaV: "Chromosomal rearrangement moves genes, but doesn’t create them. And point mutations, random as they are, don’t have any realistic chance of producing the kind of information that is necessary for protein production. Behe and Snoke make this latter point abundantly clear in their paper. So, Darwinism can invoke all the “random” mutations and rearrangements they want, but until such time as a plausible mechanism for information generation is presented, it will remain what it currently is: vacuous." I have often heard "anti-evolutionists" claim that there is no way that evolution can create information. What does that actually mean? Do you have some definition of information with which I am not familiar? Note also that there are more mechanisms creating mutations than simple point mutations and chromosomal rearrangements. Gene duplications and deletions, the oft mentioned transposons and insertion sequences, homologous recombination and integrons/gene cassettes (things I can think of off the top of my head). In what sense do these mechanisms not create information? Hawks

Ofro: "But I see no conceptual difference between a randomly occurring point mutation and a randomly occurring chromosomal rearrangement. I couldn't disagree with you more. The Modern Synthesis is based on the mathematics of Fisher, Haldane and Wright, none of which included the notion of chromosomal rearrangements and such. Bateson, at the beginning of the 20th Century, said that the fundamental challenge to evolution, in light of the newly discoverd Mendelian genetics and the segregation of alleles, was to answer the question as to how these alleles=genes came about in the first place. That is what Fisher and others applied themselves to mathematically, and what Morgan applied himself to in the lab using flies. Darwinism still has no answer to this question. Chromosomal rearrangement moves genes, but doesn't create them. And point mutations, random as they are, don't have any realistic chance of producing the kind of information that is necessary for protein production. Behe and Snoke make this latter point abundantly clear in their paper. So, Darwinism can invoke all the "random" mutations and rearrangements they want, but until such time as a plausible mechanism for information generation is presented, it will remain what it currently is: vacuous. Ofro: "The problem I have with any telic mechanism, as I pointed out earlier, is the lack of a mechanism that could direct a mutation. The information that is needed to “know” the consequence of each possible mutation is simply way too large to be stored away somewhere in the genome." I again disagree. As I mentioned in my last post, mutations have a way of canceling out the effect produced by previous mutations. Scienc is on the way of demonstrating in a clear fashion that each molecule, each atom, contains quantum information, and that molecules themselves can act as almost a computer, adjusting its information content based on the chemical/physical changes that occur to them. If this is indeed the case, then any mutation that occurs, changing the information structure of the molecule. Within this kind of a scenario, it would be possible to program a response to a "critical" (and this is sort of the operative word here) change that might occur in "critical" molecules (notably proteins). I say "critical" change because, as we know, there are so-called "neutral mutations." What I'm saying is basically that one day, soon, they will be using molecules to do quantum computations. PaV

Pav, Now you are moving away from the original purpose of your thread. The paper your presented was supposed to be an example of purposeful, directed mutations. Upon closer examination, it turned out that the mutations that were found were readily explained by a classical random-mutation-followed-by-selection scheme. There was no need to invoke some kind of intelligent scheme that targeted a gene for mutagenesis. While evolution is certainly a background theme here, it wasn’t specifically brought up here, especially with the purpose of discussing the possible difference between microevolution and macroevolution or the generation of new species. Your argument smells a bit like a distraction maneuver. I only brought in the E-word in response to the laudably logical supposition of Sladjo that all that random mutation stuff in nature should cause all kinds of genetic diseases. After all, Sladjo was quite correct with his conclusion. Now, in response to your shift in emphasis, I think we both agree that a major disagreement lies in the relationship between micro- and macroevolution. It appears that you can agree with a possible randomness in micro-evolutionary mechanisms, and if the term adaptation makes you feel better in presenting it, I’ll go along. Where the “classic” Darwin mechanism with its many small-step mutations differs from the modern version of evolution is that the palette of possible mutations is much richer now, spanning from point mutation to chromosomal rearrangement. But I see no conceptual difference between a randomly occurring point mutation and a randomly occurring chromosomal rearrangement. All of these will be subject to selection (and amplification) or deselection (and elimination) by the environmental influence. You call that adaptation, and I agree with it. Where I think we depart is my assertion that, if a sufficient number of adaptation steps occur, this eventually amounts to enough change that we can call it macroevolution. Mind you, the steps can be small (point mutation) or much more drastic (gene rearrangement) in their genomic scope and their anatomic/physiological consequence. The problem I have with any telic mechanism, as I pointed out earlier, is the lack of a mechanism that could direct a mutation. The information that is needed to “know” the consequence of each possible mutation is simply way too large to be stored away somewhere in the genome. ofro

Ofro wrote: "This is one of the reasons why most scientists are looking at mutations as a non-directed process. If an organism “knew” what to do in response to an environmental change, would it choose to pick a mutation that will kill its progeny? Evolution is a very painful process, with a huge number of failures and victims among the very few cases where a mutation results in an advantage for the offspring." I think we need to look a little closer at what happened in the experiment. What we see are different missense mutations occuring, not all at the same site, giving rise to the activation of the Gal80 gene. There have been other experiments run on bacterial mutations, missense mutations, that have concluded that these mutations tend to cancel out the effect of the others. All of this is said to point out that we're not actually seeing 'evolution' in this experiment, we're simply seeing mutational change bring about a change in gene regulation. This change in gene regulation is what is beneficial to the organism. But, what aren't we seeing? Well, we arent' seeing the formation of a new gene for metabolizing galactose--it's already there. We're not seeing a change in the galactose metabolizing gene--it remains unchanged; all that is changed is the 'off-switch', if you will. We're not seeing any morphological change. We're not seeing 'information' being added to the system. So, how can we call it 'evolution'? What we're seeing is 'adaptation'. We are seeing random mutations and artifical selection (changing the nutrient from glucose to galactose is being done humanly). So, random mutation and selection (artificial, and probably natural) results in 'adaptation', but not 'evolution'. You can call it 'microevolution', some do, but, again, what has 'evolved'? Things have changed--but, what has 'evolved'? I've said for a long time now, if Darwin had entitled his book, "Origin of Adaptations", I would be four-square behind him. In this debate with Darwinists, I think 'adaptation' should not be ceded to the Darwinists as 'evolution'. And this experiment does nothing to change my position. PaV

Joseph: "EAM has the organism initiating the mutations to find a solution to the problem- the problem being environmental changes and how to cope…" If I understand it correctly, EAM means that there is a mechanism that initiates mutations in respomse to some stimulus, as opposed to the evolutionary mechanism by which mutations happen spontaneously and continuously without a stimulus being present. So there is a principal distinction between the two mechanisms. The question is about next step, in other words, how is the mutation evaluated? How is the mutation a solution to the problem? I can't see how a cell could do that theoretically; there is no way for the information about every possible genomic mutation being stored somewhere in the cell. The only way I can see is by going through with the mutation and testing its consequence: if the mutation is beneficial, it helps the future generations, if it is neutral, it has not consequence, and if it is bad, the cell suffers. In my eyes this is the same as the mechanism of natural selection. So the only difference I can see is how the mutation comes about. Do I seee that correctly? ofro

Sladjo, “I believe that we have to clarify first what each part of the DNA is responsible for, and we have to be very sure about the existence of “non-coding” DNA or “junk-DNA” or “non-useful/non-effective” mutations…” To a large extent, the “junk DNA” argument floating around in the ID circles is a red herring used to confuse the non-biochemists among the flock of ID followers. We know the origin of much of this DNA (DNA segments that are able to move around in the genome, gene duplication etc.) but are not yet as sure about what its functions could be. However, this plays no role here, since the genome of this yeast strain is much simpler in that it has virtually no junk DNA. “On the other hand, I read (I think here, on this blog) that a DNA replication during cell division is a quite precise process, and an error (mutation) will occur only after “n” billion replications (cell divisions).” I don’t know the current state of wisdom on mutational rates in yeast, so let’s use PaV’s numbers. PaV says: “The mutation rate of eukaryotes is around 10^-8 errors/nucleotide base. So, roughly, every time it replicates ( and since its genome size is 10^8), 1 error (mutation) occurs.” PaV goes on and claims that to change an amino acid, there need to be two mutations, which is not correct. In most cases, only one nucleotide of the triplet codon that encodes an amino acid needs to be mutated to mutate the amino acid, and in our case this was verified by DNA sequencing. So if there were 10^9 cells, there is, on average, a very good chance that there is a cell in the population with a mutation in just about every position in the genome. (see: www.ucl.ac.uk/~ucbhjow/b241/mutation.html if you want to read more about the different kinds of mutations). Again, using PaV’s numbers, we can estimate that 10^9 cells take up about one milliliter (or cubic-centimeter, 1/30 of an ounce) so that a culture flask can easily hold that many cells. During the protocol, the cells were permitted to multiply, and after each round of selection most of the cells were thrown away to start the next round with a manageable number of cells. “(An extremely small error rate should be logical & predictable, otherwise I believe we would have a lot of genetic diseases, all over the world – the higher the mutation rate, the higher probability for dangerous mutations).” And that is, indeed the case. Many of the really “bad” mutations result in death before or right after birth (I urge you to look at www.kumc.edu/gec/prof/prevalnc.html to get a better feeling for how many unfortunate people are affected). Among the genetic disorders where humans can live for some time, I am sure you will recognize Sickle-Cell Disease, Hemophilia, Phenylketonuria, Cystic Fibrosis, Tay Sachs Disease, colorblindness. The list of genetic diseases is not 4000 long (en.wikipedia.org/wiki/Genetic_disorder) and growing. Many genetic disorders were actually amplified in the human population because they conferred selective advantage under some conditions: sickle cell disease protects against malaria and cystic fibrosis against diarrheal diseases; these are cases of mutations followed by natural selection we don’t like to think of. This is one of the reasons why most scientists are looking at mutations as a non-directed process. If an organism “knew” what to do in response to an environmental change, would it choose to pick a mutation that will kill its progeny? Evolution is a very painful process, with a huge number of failures and victims among the very few cases where a mutation results in an advantage for the offspring. Unfortunately, we humans are experiencing it in a very painful manner. I am sure that among the readers of this blog there are many who will agree from personal experience: as many as 1 in 25 individuals in some populations carry a defective cystic fibrosis gene (/www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=9003498&dopt=Abstract). ofro

Ofro on EAM: The link you gave is extremely vague and more of a philosophical statement than one of science, and I didn’t have the time to work my way through all the follow-up links. One cannot get any more vague than random mutations culled by natural selection. EAM has the organism initiating the mutations to find a solution to the problem- the problem being environmental changes and how to cope... Joseph

Ofro, David (vun Kannon), Thank you for your comments. However, even if your explanations are logical and they make sense, I still have a problem with this “mutational randomness” that can improve the cells capability of survival. I believe that we have to clarify first what each part of the DNA is responsible for, and we have to be very sure about the existence of “non-coding” DNA or “junk-DNA” or “non-useful/non-effective” mutations… On the other hand, I read (I think here, on this blog) that a DNA replication during cell division is a quite precise process, and an error (mutation) will occur only after “n” billion replications (cell divisions). Correct me if I’ve got it wrong. (An extremely small error rate should be logical & predictable, otherwise I believe we would have a lot of genetic diseases, all over the world – the higher the mutation rate, the higher probability for dangerous mutations). That’s why I have a problem with this continuous “mutation process”… And that’s why I believe (making an analogy to the man-made adaptive systems) that there is a “pattern” that will be applied by a kind of “computational unit” probably present somewhere in the cell, when an environmental change is detected. Because if we translate the discussion to the multi-cellular world (i.e. - Darwin’s finches), we may have a big problem… Sladjo

Hawks wrote: "Also, reading DaveScot’s post, I did not see him claim that the methylation process was anyhow directed (I might have missed this, of course). There's an article in this week's New Scientist about methylation control. You can't read the whole article unless you have an online subscription to the magazine, but the portion of the article you can read is sufficiently provocative along these lines. Here's the link: http://www.newscientist.com/channel/life/mg19125685.500-genome-onoff-switches-mapped.html PaV

PaV wrote: "Mathematically, it appears we can\’t \’rule out\’ a truly \”random\” source of mutation. However, I duly note Dave Scot\’s post and his private work with cultures. He doesn\’t rule out a \’directed\’ methylation process." There are a lot of known processes - and potentially a lot of unknown ones - that can create changes in the phenotype of an organism. There is no absolute reason why any of these should be ruled out. In the absense of any positive evidence FOR these processes there is, however, no scientific reason to claim that they are responsible either. Also, reading DaveScot's post, I did not see him claim that the methylation process was anyhow directed (I might have missed this, of course). Hawks

DaveScot, I agree that epigenetic, non-Mendelian genetic phenomena can play a role in some aspects of inheritance. I like the fact that there are proposed biochemical mechanisms and experimental evidence. What I don’t know is to which extent this is also a mechanism for adaptation, but it could be. Neither could I comment on how an epigenetic effect becomes fixed. I suspect that this is related to the mechanisms that maintain a cell type even through cell divisions. In our specific case here, there was is a way to test for genetic vs. epigenetic: introduce the identified mutation into a wild-type cell and test whether the mutant phenotype is artificially re-created. And that is what the researchers did to confirm that they isolated specific point mutants. (And with respect to your experiments, you may have found an epigenetic phenomenon, but your cells may well have been a mutation to begin with. BTW: keep you eye open for possible roles of catalase and superoxide dismutase (if your organisms have them) and their possible effects on cell vigor and longevity; there may be some literature on that). Joseph: I am much more ambivalent about the notion of endogenous adaptive mutagenesis, unless the epigenetic mechanisms mentioned by DaveScot and EAM are considered equivalent. I want to know more about possible mechanistic aspects of EAM before I buy into that. The link you gave is extremely vague and more of a philosophical statement than one of science, and I didn’t have the time to work my way through all the follow-up links. For example: “EAM requires that organisms use these mental phenomena to actively attempt to 'learn' to adapt, by means of a trial and error heuristic experience in which a 'best available solution' is sought to a specific 'problem'.” Come to think of it, this almost sounds like he is proposing that the cells/organisms do a little random mutation here and there, followed by selecting ;) PaV, next time I'll be precise and stay with "millions" and "billions". I did not mean to imply 2^700 cells. ofro

Well, I\'ve had time to look at some of the posts and to look at the paper. I see, as has been pointed out by a number of you, that I was working with a wrong notion of missense mutations. One point mutation a missense mutation does, indeed, make. The confusion came from another paper I had looked at a while back. This does, radically, change the numbers, so that, for example, in the last post I put up, to get a point mutation/misense mutation \"randomly\" would require only 45 x 10^8 cells, which represent a cubic volume with 18 inches per side. This is very realistic. So, indeed, these missense mutations can be gotten at \"randomly\" in a realistic fashion. I have to simply cede that point and do some re-thinking myself. Thanks to those that have pointed out my error. Mathematically, it appears we can\'t \'rule out\' a truly \"random\" source of mutation. However, I duly note Dave Scot\'s post and his private work with cultures. He doesn\'t rule out a \'directed\' methylation process. (P.S. Dave, I can\'t tell you how impressed I am that you\'re doing the kind of lab work you\'re doing. Keep it up!) PaV

Ofro wrote: "If I let the cells grow for 700 generations, there are bound to be a small number of mutations in the genome of zillions of yeast cells, and these mutations will be randomly distributed throughout the genome." Let's take a look at what these "zillions" of cells would look like. I suppose your position is that one cell dividing into two for 700 generations would produce 2^700 progeny. I guess that's the "zillions" you're talking about. But, let's look at that number a bit more closely. 2^700=10^210(approximately). The average size of a yeast cell is between 5-10 microns. For ease of calculating, we'll use the 10 micron figure, which means each cell is 10^-5 meters in diameter. This means that in a cubic meter (m^3), there would be 10^15 yeast cells found. Now, let's divide the 10^210 cells by this 10^15 cells/cubic meter. This gives 10^195 cubic meters. Dividing the exponent by 3 (to go from cubic meters to just plain old 'linear' meters) we get a length of one side of this cube of 10^200 cells of 10^65 meters. The speed of light is 3x10^5 meters/sec. Dividing the 10^65 meters by this figure gives 3.333 x 10^59 light-sec. Now there are 3.15 x 10^7 seconds in a year. So, dividing 10^59 light-sec by 3.15 x 10^7 seconds/yr gives us, roughly, 10^52 light-years of distance for the ONE SIDE of these 2^700 cells. The universe is considered 13, or so, billion of years old. That means light coming to us from the Big Bang is 13 x 10^9 light-years away. That means that these 2^700 yeast cells would create a cubic volume that is 10^39 (roughly) times the size of the visible universe. I think you see my point. You need to re-think your "zillions" of cells. The genome size of yeast cells is around 45 million base pairs (Mbp). A random mutation has a one in 45 million chance of happening. But, missense mutations have at a minimum TWO mutations. So now you need to have one mutation happen at just the right place in Gal80 gene, and then you need another mutation adjacent to the first mutation. And each of these mutations have a one in four chance of being the right one since, after all, there's four nucleotide bases to choose from. So the odds of this "minimal" missense mutation happening right there in the Gal80 gene is: (roughly) (2 x 10^-8)x(2x10^-8)x(.25)x(.25)=2.5 x 10^-17. So, if you have 2.5 x 10^17 cells, then you cold expect this "minimal" missense mutation to occur "randomly". Now, what do these 2.5 x 10^17 represent? Well, based on my prior calculation, there are 10^15 cells/cubic meter. So we're looking at enough yeast cells to fill a volume that is 2,500 meters on a side; that is, a volume that would extend from Los Angeles to Dallas, Texas on one side, from Los Angeles and Dallas up past the Canadian border, and sticking 1,500 miles up into the air. Do you really think that we're dealing with "random" mutations now? And what if the missense mutation was minimally 3 nucleotides long, then what? And, of course, some of these strains developed more than one missense mutation in the Gal80 gene. That means we might have to double, triple, quadruple the volume of yeast cells we need to "randomly" arrive at these mutations. What do you think of all of this Ofro? PaV

Wow, a truly exciting thread on UD! PaV, this is a highly intriguing find. bFast

Ofro, It isn't that I am worried about transposons, the point being made is that we declare "random(ness)" before we actually understand it. That said your point about the single-nuc substitutions are duly noted and it appears to meet Dr. Spetner's criteria. However it could also point to an easily manipulated design (as in ID) feature for survival. And then there is EAM- Endogenous Adaptive Mutagenesis Joseph

Sladjo wrote: My logic (maybe be wrong, nevertheless) says to me that the yeast somehow DETECTED the environmental change and somehow triggered some VERY SPECIFIC type of mutation, and this had an immediate effect on adapting the bug for this new environment. I can't comment on whether yeast can sense the concentration of nutrients in their environment, and thereby improve their metabolism. But that is not the core issue here. Say you had a trillion yeast in a culture, reproducing every twenty minutes. If they make mistakes at the rates given by PaV, then every twenty minutes there will be about ten million individual yeast that have picked up a mutation in a given region that is a thousand bases long. These ten million individuals now have a better chance of surviving in the culture, because they have a better match to the available nutrient mix. While these ten million got lucky, the other 999,990,000,000 yeast mutated in some way that was bad to neutral. They starved for glucose while our ten million lucky friends were chowing down on galactase. So the yeast population didn't sense and detect and plan a specific mutation. With such a large number of individuals, it was testing EVERY mutation possible, every twenty minutes, which makes the population incredibly agile at adapting to a changing environment. Individual yeast don't need elaborate detection apparatus when the population can use mutation as a tool to sense changes in the environment. Random mutation is testing the environment every twenty minutes. (Obviously, saying that a population "uses" mutation is just a description of the aggregate outcome that is easy to fit into human language. There isn't any quality of purpose to the yeast population, any more than air has purpose when it flows from a high pressure container to a low pressure container through a pinhole. The aggregate description just summarizes a lot of random activity. The air and the yeast population are just following a temporary gradient.) David vun Kannon

Ofro, there is so much we don't know. Gene induction via methylation is the usual mechanism in fungi adapting to new substrates. I've got some experiments going on in my hobby lab with peroxidase gene induction in the mushroom volvariella vovacea (paddy straw mushroom). This is the mushroom commonly found in Chinese food. It has many unusual charateristics including extremely rapid metabolism, rapid senescence in pure tissue culture, and intolerance of cold temperatures. I've been experimenting with a single pure tissue culture for about a year now investigating ways to make culture maintenance of this notoriously difficult to maintain species easier. I've met with much success and along the way I was experimenting with peroxide in the agar to inhibit fungal contaminants so that less onerous sterile technique can be used in handling the cultures. Surprisingly one of the cultures survived transfer onto an agar that was 10 times the concentration any other species seem able to tolerate well. In fact the resulting culture when returned to non-peroxide agar was quite a bit more vigorous than a control culture the same number of generations removed from the original culture. After being removed from the peroxide environment for 6 weeks and God only knows how many cell divisions I transferred back onto peroxide agar. The results were ambiguous and I'm still experimenting, but it appears the adaptation wanes over time but persists to some degree for as long as I've been doing it (about 16 weeks now). Anyhow, I wouldn't be surprised if epigenetic fast-track reversible evolution via methylation had a mechanism on top of it that caused the epigenetic modifications to become more permanently fixed as genetic modifications. DaveScot

Joseph, In this particular instance there is no need to worry about transposon-related mutations since we know that these were single-nucleotide substitutions that resulted in an altered codon and a change to a different amino acid. Everything is readily explained with random point mutations, followed by a mechanism that permits just one class of mutants to grow and multiply faster and take over the population of cells, and then get characterized experimentally as to precisely what the mutation was. So these experiments are actually a picture-book case for RM&NS. ofro

Sladjo, “I read all the posts, and I didn’t find any clue about WHAT triggered the mutations… I mean HOW the yeast’s DNA “knew” that there is a need for mutations…” There is nothing in particular that triggered the mutations that were found. During every cell division when the genomic DNA is replicated, there is a chance of errors being made (and there are plenty of other occasions when DNA can be damaged and there is an error in the repair mechanism). These mutations occur more or less randomly. Most of these mutations are in places where it doesn’t matter much for the cells. A few mutations did matter here because an artificial environmental stress was imposed that selected the yeast cells that were mutated such that they could more rapidly respond to a switch to galactose as sole energy source. As a consequence they could grow and multiply faster, and after 700 generations there is a good chance that the population of cells is greatly enriched in these mutants. So the cells or their DNA does not need to “know” if there was a change in the environment. (to be precise: the cells have a mechanism to detect the presence of glucose or galactose, which is what turns one pathway on and the other off; see my response to DaveScot. However, this is not what you mean here; I presume you are referring to the cells knowing if they need to mutate.) The situation is more like a lottery. The cells that were “lucky” to have the accident called “beneficial mutation” have a better chance to grow. They didn’t know that beforehand when the mutation occurred (in them or their parent cells). But the overall result is that they won the lottery. The reason why the experimental result appears to be specific to Gal80 and directed is that ONLY the lottery winners (the Gal80 mutations) were able to grow faster, not all the other “ticket owners” that had acquired a mutation in a different gene that had no influence on how they grew on galactose. ofro

Random mutations? My guess is that when one really doesn't understand what is going on it is easy to declare "random" and move one- Dr. Spetner discussing transposons:

The motion of these genetic elements to produce the above mutations has been found to a complex process and we probably haven’t yet discovered all the complexity. But because no one knows why they occur, many geneticists have assumed they occur only by chance. I find it hard to believe that a process as precise and well controlled as the transposition of genetic elements happens only by chance. Some scientists tend to call a mechanism random before we learn what it really does. If the source of the variation for evolution were point mutations, we could say the variation is random. But if the source of the variation is the complex process of transposition, then there is no justification for saying that evolution is based on random events.

Joseph

moderator: looks like another comment I sent is stuck somewhere in the queue, or perhaps in the spam filer? Indeed. No longer. -moderator ofro

DaveScot: “Where I think ofro is wrong is that there was any random mutatoin based evolution involved. Fungi adapt to different food sources and they don’t do it by accident. If a gene actually changed sequence to go from glucose to galactose I’d be looking for a mechanism whereby that mutation was directed. It was not likely serendipity that did it.” I think it boils down to what you mean with “directed” and “serendipitous”. But before I get to that, it is important to distinguish between adaptive mechanism and genetic changes in response to an altered environment. The cells certainly have built-in adaptive mechanisms. The regulatory mechanisms that turn/off the proteins metabolizing glucose and those that turn on/off galactose metabolism are such mechanism ( and these two regulatory mechanisms share some components, I believe). That is why the cells could be switched from one substrate to the other. The explanation for why these mechanisms exist at all is related to the cell’s overall energy cost: in the long run, it is energetically more economical to shut off a certain metabolic pathway if there is no substrate around as opposed to synthesizing the proteins and maintaining that pathway. That is different from genetically adapting to a new environmental situation, as imposed here experimentally. The result was a change of the properties of one of the regulatory proteins such that it didn’t shut off the galactose-utilizing pathway, or at least not as rapidly, so that it could be re-activated more readily. As I explained in my response to PaV, this mutation was, from what we have known about the regulatory pathways involving glucose and galactose utilization, an experimental expectation. (I think the authors were actually not out to demonstrate that but were looking for a different phenomenon that I haven’t tried to understand yet). Back to directed and serendipitous. It seems to me that you are looking for some sort of telic mechanism that predisposes the yeast cell to change its DNA sequence in response to an altered environment, beyond the RM&NS paradigm. The reason why I don’t think such a mechanism exists, or at least is not necessary, is the following. At the end of the cycling protocol, when a mutated cell was isolated from each of the four batches, there were three different strains with different galactose phenotypes (of the four, 2 were identical). What would happen if we put them back to a steady diet of glucose when they don’t need to handle galactose? My interpretation of your notion is that you would expect them eventually to switch back genetically to the original DNA sequence, or perhaps just stay the same genetically. The latter case will, in the long run give them an energetic disadvantage because now they are keeping the galactose-utilizing pathway on without getting anything out of it. The former case is unlikely. In microbiology, there is a long history of looking for mutants in a certain phenotype and subsequently for revertants, i.e. new mutants in which the original phenotype was restored. In the great majority of the cases, the mutation in the revertants was not restored, but instead a different mutation was found that changed the protein’s function to nearly the same as the original wild-type. The reason is that there are many ways to restore a protein’s function compared to the one way to return to the original condition. It is simple statistics. In the same way, there is a chance that, if you wait long enough, there will be a revertant with a galactose phenotype similar to the original one, but it will most likely not be identical to the original strain. According to your directed mutation scheme, I would expect a return to the original genetic state. If I misunderstood it, and you would not make such a prediction, then there is no experimental way to distinguish between the two models, and I would prefer the simpler mechanistic description I outlined above. ofro

PaV: (sorry if this post is going to be lengthy) It really depends on what you consider as random. The results of the study show that the best way for the yeast to switch faster to a galactose-utilizing metabolic pathway was to mutate the Gal80 protein, and to mutate it only in a certain ways. In that sense the mutation wasn’t random. However, that does not mean that the mutations were non-random in the sense of being targeted. If I let the cells grow for 700 generations, there are bound to be a small number of mutations in the genome of zillions of yeast cells, and these mutations will be randomly distributed throughout the genome (or at least close to random; there could be – due to some special DNA-structural properties – some regions with a higher and some regions with a lower mutational probability). But we would never detect these random mutations with the protocol used, since at the same time the cells were subjected to a nutrient-switching protocol. The beauty of natural selection is that only mutations that permit the cells to grow and multiply faster will have an advantage over the other cells. Damaging mutations will either kill or impair cells, and neutral mutations won’t have a consequence on the ability to switch to galactose. So among all the random mutations that occurred, this artificially imposed environmental stress selected, by necessity, those that gave the cells a selective advantage. That is why the results give the appearance of non-randomness. There is an additional level of random vs. non-random. We know now that the major advantage for a cell prospering in this protocol is achieved if a certain protein (called Gal80) is mutated. Are the mutations randomly distributed within this protein? This is preliminarily answered by asking: could a mutation anywhere in this protein give a selective advantage? The answer is no; if a mutation were neutral, there would be no selection. The only mutations that could be detected are those that downregulate the protein activity (and therefore de-repress the galactose pathway). And indeed, the experimental observation is that among the four independently raised strains there are three different sets of mutations in “conserved sites.” The implication of “conserved” is that among all the analogous Gal80 proteins found in other yeast strains that can differ in their sequence by as much as 60%, these amino acids are always the same because changing them would destroy function. So you would expect mutations in such critical sites. Two mutants have the same mutation in a single amino acid, and the other two are double-mutants where two amino acids were changed. Four of these amino acid sites were known from other experiments to weaken the repressor function of Gal80, one directly and three by not being able to go the nucleus where the protein is supposed to bind to DNA. So even though the results appear non-random to the casual observer, they were random. The five critical amino acid locations detected here were random selections from among the number of possible critical sites (I don’t know the exact number). Five of them were found, perhaps more would be found with an extended search. ofro

The (highly scientific) language used in the article and in some of your posts exceeds my knowledge, so it may be I didn’t understood all that genetic stuff, but… I read all the posts, and I didn’t find any clue about WHAT triggered the mutations… I mean HOW the yeast’s DNA “knew” that there is a need for mutations… My logic (maybe be wrong, nevertheless) says to me that the yeast somehow DETECTED the environmental change and somehow triggered some VERY SPECIFIC type of mutation, and this had an immediate effect on adapting the bug for this new environment… Please correct me if I’m wrong… Sladjo

HodorH "I think it’s misleading to call it directed. Accelerated would be more clear." I think it's misleading to call it random mutation. Induced mutation would be more clear. In fact the researcher at Scripps who discovered the LexA pathway in e.coli calls it induced mutation. http://biology.plosjournals.org/perlserv/?request=get-document&doi=10%2E1371%2Fjournal%2Epbio%2E0030176 DaveScot

PaV wrote: "And, for a missense mutation, a minimum of two nucleotide errors is needed. And, these mutations have to occur next to each other....... So, tell me, how did these mutations–all in the same GAL80 gene–come about ‘randomly’?" I'm not sure why you claim that a missense mutation requires two "nucleotide errors". It doesn't. And, in fact, the mutations listed in the four strains only required one mutation each to make missense mutations. The rest of your calculation assume that the specific mutation would occur in ONE yeast cell. There are preciously few experiments done using a single microorganism. A typical experiment would contain millions if not billions of cells. I don't think your numbers in post #8 bear out the fact that we’re not dealing with anything random here. Hawks

PaV - where do you get the idea that two mutational events are needed to get a mis-sense mutation? I would have expected that they would most commonly arise when the wrong nucleotide is added at replication, which would be a single event. Bob Bob OH

PaV writes:

Michael Denton would likely agree with you. But he takes a Platonic view of all of this and takes the position that all of creation is “optimal”, and therefore directed by the “natural laws “ (with the presumption that God is responsible for those laws) to their rightful “end”. But I have a feeling you wouldn’t go along with him here. Am I right?

No, I'd say I agree with Denton more or less on this point. Almost. All of creation is optimal for fulfilling the Creator's purposes except those with free will (probably only humans).

HodorH further wrote: That said, if you’re telling me that these positive mutations occurred at a higher rate where they became beneficial to the yeasts than elsewhere in the genome, then I’m telling you to go back to the lab and do that experiment. Sorry, but I’m not getting the point you’re trying to make. Perhaps you can rephrase it.

Without good evidence, we should assume that mutations occur in this gene at the same frequency as in other genes. This paper offers no such evidence.

That’s not quite accurate. The mutations occurred in the same gene (GAL80), but the number of missense mutations differed between the four strains.

Two of the four strains had the same mutation. The other two had 2 mutations each, all different from the shared mutation from the other two.

An environmental stress is imposed on each of these four strains of yeast. Each of these strains developed a missense mutation(s) in the GAL80 gene. The article says that were dealing with an area of about 1kb, i.e., 1000 nucleotide bases. Now the length of the yeast genome is probably 10^8 nucleotides long. The 1kb therefore represents 10^3 nucleotide bases. The mutation rate of eukaryotes is around 10^-8 errors/nucleotide base. So, roughly, every time it replicates ( and since its genome size is 10^8), 1 error (mutation) occurs. And, for a missense mutation, a minimum of two nucleotide errors is needed. And, these mutations have to occur next to each other.

I'm afraid this isn't right at all. Missense mutations only require 1 base pair to change. Where did you get this idea? It would be a good idea to track down the source of this misconception; basic errors are the last thing we need.

So, the odds of getting ‘one’ mutation in the 1kb region, that is, the ‘first’ mutation (error), will take, on average, at least 10^5 generations of the yeast (and thus surpass the number of possible nucleotide sites that lay ‘outside’ the 1kb region). Now, to get the ’second’ mutation, it isn’t sufficient for it to lie in the 1kb region; it needs to be ‘next’ to the ‘first’ mutation. So, if the genome size is 10^8 bases, and the error rate is 10^-8 errors/base, then, again, we get 1 error/replication (generation). So now we need 10^8 generations for the ’second’ mutation to occur since there is only one place in 10^8 locations that lie ‘next’ to the ‘first’ mutation). If we assume that the generation time of yeast is 20 minutes, 10^5 generations represents (10^5 generations x .333 hrs/generations)/24 hrs/day/365 days/yr=3.8 years. 10^8 generations is a thousand times more; so it represent 3800 years. So, tell me, how did these mutations–all in the same GAL80 gene–come about ‘randomly’? Has the lab been in existence for 3800 years?

Hmm, you seem to be doing calculations for a constant population size of 1. My above objection obviously applies to this section as well. HodorH

DaScot says

Regardless, directing mutations at a few genes vs. a large number is the difference between hunting with a rifle and hunting with a shotgun. Both are still directed methods.

More like an atomic bomb. There's no evidence from either paper that ANY specific genes or chromosomal regions were targetted. I think it's misleading to call it directed. Accelerated would be more clear. PaV, my response is forthcoming, but may take a bit longer. HodorH

hodor "Increasing the rate of mutation is not the same as directing the location of mutations." I don't entirely agree. I doubt the mechanism in question accelerates mutations in critical sequences like those that define the genetic code. That's certain death. So there is probably at least some very rough targeting. Regardless, directing mutations at a few genes vs. a large number is the difference between hunting with a rifle and hunting with a shotgun. Both are still directed methods. Turning up the mutation rate in response to the environment is, in and of itself, directed mutation. DaveScot

To Hawks, and ofro, and HodorH: please read my last post. I think it's obvious that, as I stated in the post, MUTATION has happened, but the question is whether it was 'random' or not. I think my numbers in the last post bear out the fact that we're not dealing with anything random here. But if I'm wrong, then please point out where I have erred. PaV

HodorH wrote: All mutations are directed. They are directed by the laws of the created universe, in which their organisms reside. Michael Denton would likely agree with you. But he takes a Platonic view of all of this and takes the position that all of creation is "optimal", and therefore directed by the "natural laws “ (with the presumption that God is responsible for those laws) to their rightful "end". But I have a feeling you wouldn't go along with him here. Am I right? And, of course, you know full well what I meant by 'directed', so why the need for equivocation? HodorH further wrote: That said, if you’re telling me that these positive mutations occurred at a higher rate where they became beneficial to the yeasts than elsewhere in the genome, then I’m telling you to go back to the lab and do that experiment. Sorry, but I'm not getting the point you're trying to make. Perhaps you can rephrase it. And lastly HodorH wrote: You should also note that not all 4 strains received the same mutation. That's not quite accurate. The mutations occurred in the same gene (GAL80), but the number of missense mutations differed between the four strains. Let me, however, just verbally clarify things here: An environmental stress is imposed on each of these four strains of yeast. Each of these strains developed a missense mutation(s) in the GAL80 gene. The article says that were dealing with an area of about 1kb, i.e., 1000 nucleotide bases. Now the length of the yeast genome is probably 10^8 nucleotides long. The 1kb therefore represents 10^3 nucleotide bases. The mutation rate of eukaryotes is around 10^-8 errors/nucleotide base. So, roughly, every time it replicates ( and since its genome size is 10^8), 1 error (mutation) occurs. And, for a missense mutation, a minimum of two nucleotide errors is needed. And, these mutations have to occur next to each other. So, the odds of getting 'one' mutation in the 1kb region, that is, the 'first' mutation (error), will take, on average, at least 10^5 generations of the yeast (and thus surpass the number of possible nucleotide sites that lay ‘outside’ the 1kb region). Now, to get the 'second' mutation, it isn't sufficient for it to lie in the 1kb region; it needs to be 'next' to the 'first' mutation. So, if the genome size is 10^8 bases, and the error rate is 10^-8 errors/base, then, again, we get 1 error/replication (generation). So now we need 10^8 generations for the 'second' mutation to occur since there is only one place in 10^8 locations that lie ‘next’ to the ‘first’ mutation). If we assume that the generation time of yeast is 20 minutes, 10^5 generations represents (10^5 generations x .333 hrs/generations)/24 hrs/day/365 days/yr=3.8 years. 10^8 generations is a thousand times more; so it represent 3800 years. So, tell me, how did these mutations--all in the same GAL80 gene--come about 'randomly'? Has the lab been in existence for 3800 years? I think you see my point here. I welcome your response. PaV

Ofro is sort of right. Fungi are masters at optimizing digestive enzymes for the substrate they find themselves on or in and they have a large repository from which to choose the kind and amount of enzymes. Labs where commercial strains of mushrooms are kept in culture know this and it's common practice to both periodically change the agar recipe so they don't become adapted to potato dextrose and stop producing enzymes to digest malt extract (for instance) and to put a small amount of their fruiting substrate into the agar (sawdust or straw for example) so they continue producing the proper enzymes to digest those lignin rich nutrient sources. Once they've adapted to any particular nutrient source it can take quite a while for them to adapt to something else. Where I think ofro is wrong is that there was any random mutatoin based evolution involved. Fungi adapt to different food sources and they don't do it by accident. If a gene actually changed sequence to go from glucose to galactose I'd be looking for a mechanism whereby that mutation was directed. It was not likely serendipity that did it. DaveScot

PaV: I think you fell prey to a thorough misunderstanding of the experiments. You are interpreting the fact that all four “evolved” yeast strains had a mutation in the same gene was due to some form of directed, non-random event. The reason why the mutation occurred in one gene is one of experimental necessity. Just look at the conditions to which these yeast strains were conditioned. They were subjected to a protocol involving 36 cyclical changes of the growth media between containing either glucose or galactose as energy source. To go from metabolizing one sugar to metabolizing the other, the yeast has to turn on one metabolic pathway and turn off the other. Looking for a phenotype that exhibited a faster metabolic switch to galactose, the researchers found that all four strains had a mutation in a gene that regulates the expression of the galactose utilization pathway. (It normally suppresses the galactose pathway, and the mutations reduced the repressor activity). Now let’s ask where you would expect a beneficial mutation to occur. Take a random pick among the thousands of yeast genes. Would you expect that a mutation in a DNA repair or a drug resistance or an amino acid metabolic mechanism have a selectable effect on the galactose response? Or would it be more reasonable to expect a mutation in one of the regulatory proteins involved in turning on the galactose pathway? Nothing mystical about it. ofro

If e.coli, a lowly bacteria, can control mutation rate in certain genes in response to something in the environment you should be reluctant to deny the same capability to other organisms which are equally or more complex without evidence.

Unless I misread, those mutations are not directed. Increasing the rate of mutation is not the same as directing the location of mutations. I also would expect bacteria to have more adaptable genomes, as (some of them) they persist in varied enviroments, and thus must be able to adapt quickly. HodorH

Hodor - read this before you further dismiss directed mutations. If e.coli, a lowly bacteria, can control mutation rate in certain genes in response to something in the environment you should be reluctant to deny the same capability to other organisms which are equally or more complex without evidence. DaveScot

PaV wrote: "Think about it: ALL four ‘evolved’ strains basically hit on the same mechanism. We certainly have change (mutation), but is it ‘random’ if each of the four strains reacts in the same way? How probable is it for a mutation to occur in the same place in all four strains while causing the same changed metabolic pathway to be set in motion? Random mutation? I think not." The article in question says also: "We then applied this method to four yeast strains that had independently adapted to a fluctuating glucose–galactose environment." Since the adapted yeasts strains had already gone through SELECTION for the trait desired, it would hardly be surprising if any mutations discovered were not random. Your question "We certainly have change (mutation), but is it ‘random’ if each of the four strains reacts in the same way?" should really be restated as "Given that the four strains react in the same way, should we expect any mutations found to be random?" The answer is no. Hawks

This could be (a definite could be) evidence for Dr Spetner's "built-in responses to environmental cues". Perhaps "Not By Chance" needs to be revisited... Joseph

All mutations are directed. They are directed by the laws of the created universe, in which their organisms reside. That said, if you're telling me that these positive mutations occurred at a higher rate where they became beneficial to the yeasts than elsewhere in the genome, then I'm telling you to go back to the lab and do that experiment. You should also note that not all 4 strains received the same mutation. HodorH