Average child has 60 genetic mutations?

_{Denyse O'Leary

June 13, 2011

Darwinism, Genetics, Intelligent Design}

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

From “We Are All Mutants: First Direct Whole-Genome Measure of Human Mutation Predicts 60 New Mutations in Each of Us,” (ScienceDaily, June 12, 2011), a study involving four adults and one child, we learn:

Each one of us receives approximately 60 new mutations in our genome from our parents. This striking value is reported in the first-ever direct measure of new mutations coming from mother and father in whole human genomes.[ … ]

Mutations that occur in sperm or egg cells will be ‘new’ mutations not seen in our parents.
Although most of our variety comes from reshuffling of genes from our parents, new mutations are the ultimate source from which new variation is drawn. Finding new mutations is extremely technically challenging as, on average, only 1 in every 100 million letters of DNA is altered each generation.

A surprise was the considerable variation in families, as to whether most mutations arose from the father or the mother. In theory, the father was favoured as a source of mutations because of “ the additional number of times that the genome needs to be copied to make a sperm, as opposed to an egg.”

Assuming the results hold up, what would they suggest about human evolution?

Comments

Obviously they were poor, as I have said. So are you going to tackle my posts regarding the distribution of mutations along the selection coefficient axis?Elizabeth Liddle_{June 26, 2011
June
06
Jun
26
26
2011
03:27 AM
3
03
27
AM
PDT}

Now, about those distributions….
Now, about those word choices...Mung_{June 24, 2011
June
06
Jun
24
24
2011
10:27 PM
10
10
27
PM
PDT}

OK, he discounts it, then. Golly, Mung, it's like talking to a lawyer. Are you a lawyer? Still fair enough. Bad word choice on my part. Now, about those distributions....Elizabeth Liddle_{June 22, 2011
June
06
Jun
22
22
2011
12:03 AM
12
12
03
AM
PDT}

Sanford:
Epistasis - The different mutations that affect the same trait often interact, and when this happens, it is called epistasis. A deleterious mutation may be much more or less deleterious depending on the absence or presence of other mutations. Such epistasis creates non-heritable noise and strongly interferes with selection. Geneticists acknowledge that epistasis is important, but assume that positive and negative interactions largely cancel out.
Elizabeth:
Then there is the whole issue of epistasis, which Sanford ignores.
So when you said that Sanford ignores the whole issue of epistasis you didn't mean to imply that he is ignorant of epistasis? Sanford:
Synergistic epistasis - The term synergistic epistasis is normally only used in attempting to rationalize how genomes might be prevented from degenerating continuously. The basic concept is that epistasis (interaction) between mutations is consistently negative. Therefore, as mutations accumulate, each new mutation has a greater and greater average fitness deleterious effect. This is the exact opposite of the standard multiplicative population genetics model, wherein each mutation has less and less effect (one or both models must be wrong). The synergistic epistasis model is extremely artificial and biologically un-realistic. Even if the model is granted, it can be shown that this mechanism fails to stop degeneration when linkage and the interaction between mutations and non-mutations are also taken into account.
Elizabeth:
Then there is the whole issue of epistasis, which Sanford ignores.
And when you said that Sanford ignores the whole issue of epistasis you didn't mean to imply that he failed to mentioned it multiple times? Sanford:
This problem of the fundamental inter-relationship of nucleotides is called epistasis. True epstasis is essentially infinitely complex, and virtually impossible to analyze, which is why geneticists have always conveniently ignored it.
Elizabeth:
Then there is the whole issue of epistasis, which Sanford ignores.
Funny, Sanford accuses your side of ignoring it. Sanford:
synergistic epistasis .. means that mutations interact such that several mutations cause more damage collectively than would be predicted by their individual effects. At least one paper provides experimental evidence that the concept is not valid (Elena and Lenski, 1997). But even if it were valid, it makes the genetic situation worse, not better. We have always known that genic units interact, and we know that such epistasis is a huge impediment to effective selection. This fact is ignored by most geneticists because selection scenarios become hopelessly complex and unworkable unless such interactions are conveniently set aside. But now, when genetic interactions can be usedto cloud the problem of error catastrophe, the concept is conveniently brought forth and used in an extremely diffuse and vague manner, like a smoke screen
Elizabeth:
Then there is the whole issue of epistasis, which Sanford ignores.
Only if by ignore you mean addresses it head on.Mung_{June 21, 2011
June
06
Jun
21
21
2011
05:54 PM
5
05
54
PM
PDT}

Mung @ #77
Elizabeth, I am looking at Fig. 3d. The x (horizontal) axis is labeled mutation effect. Sanford has plotted a curve of decreasing frequency (the y axis) as the mutation effect becomes more positive. Yet you claim that Sanford does not do that. You are mistaken.
he takes Kimura’s theoretical model for deleterious mutations only, shows the “VSDM” zone (i.e. those theoretically deleterous mutations that are “effectively” neutral, and then jams next to the zero line on the positive side, the population of observably beneficial mutations (BMs), totally ignoring the nearly neutral zone between zero and BMs(i.e. the VSBM zone).
I see a curve. Don’t you see a curve? The frequency decreases as the effect increases. That’s what I see. Every one of those graphs has the same label for the x axis (mutation effect). Not only does he show a curve, but the entire curve falls within the “no selection zone”! So you are doubly mistaken. The text for 3d even reads: “it becomes obvious that essentially all beneficial mutations will fall within Kimura’s ‘no-selection zone’.” If you are trusting your memory, let me suggest that you actually refer to the book instead.
Mung, I have the book right by me, but perhaps I didn't make myself clear. Let me try again: Figure 3a shows a normal distribution with a mode at zero on the x axis. The x axis is labeled "mutation effect". The caption makes it clear that this is meant to represent positive fitness greater than zero and negative fitness below zero. Sanford does not say relative to what, so we must assume that it is relative to some ancestral population (as opposed to a competing population). He calls this the "naive view", as indeed it is. He then present Figure 3b, in which he simply removes the entire distribution that is greater than zero, leaving a "cliff edge" as you go cross the zero line. He says this represents the additional information that "essentially all mutations are deleterious". He does not reference this assertion. However, we can agree that in a well adapted population, the vast majority of mutations will be either neutral or deleterious. But he says this distribution is "still too optimistic" and then presents Figure 3c. In this figure he has, firstly, replaced his half-bell curve with an exponential (or similar) function, and says it is "adapted from a figure by Kimura (1979)". However, the "adaptation" consists of extrapolating Kimura's graph beyond zero and assuming that the frequency beyond this point is in fact zero - that the frequency of mutations rises astronomically as we approach zero from below, then plummets to zero. He offers no justification for this extrapolation, and does not even mention the fact that Kimura's graph does not even attempt to model the distribution values greater than zero. Kimura's graph is a mathematical convenience, it does not represent data. He also shades in a zone either side of zero which he calls a "no selection zone". In this he follows Kimura, except that of course Kimura does not model the distribution above zero. He finally presents Figure 3d, in which he adds a tiny triangle, within the "no-selection zone" at the base of his otherwise vertical cliff face at zero. He says that even his tiny triangle is enlarged so that it will show up on the drawing. So again, he is presenting a cliff face at zero, with a tiny pile of rubble at the base, not extending outside the "no selection zone". He derives the size of this triangle from, among other sources, Gerrish and Lenski (1998). He then concludes that this is a realistic distribution and reveals that even the tiny percentage of "beneficial mutations" that exist are in the "no selection zone". He has arrived at this conclusion in my view by a) piling error on error and b) ignoring actual data. His first two plots I will ignore as one is indeed "naive" and the second just silly. His third is, as I've said, a completely unsupported extrapolation from a theoretical paper by Kimura that dealt only with deleterious mutations, and was therefore able to utilise a monotonic function. And his fourth compounds the error made in his third by a misreading of Gerrish and Lenski. In fact, a data paper that he cites three times shows that his Figure 3d is wrong, in several ways, as indeed it must be. The Elena et al (1998) actually gives a distribution of mutations that shows that of course there is no spike at zero followed by a cliff edge, but a mode at rather less than zero, the curve starting its descent before zero is reached, descending more or less smoothly, and, in Elena's sample, terminating before exiting the "no selection zone". So he'd have been better off in some ways simply going with Elena's distribution, which at least makes some sense. But his real error is in misinterpreting Gerrish and Lenski (1998) who, far from computing a "small triangle" within the "no selection" zone, give their figure (albeit a small one, and remember this is an asexually reproducing population) for beneficial mutations that are actually selected (and go to fixation). In other words, taking Elena et al and Gerrish & Lenski together, we must conclude that in asexually reproducing populations of E Coli, most mutations are deleterious compared with the ancestral population, some very deleterious, and those that are beneficial compared with the ancestral population, only a minor proportion go to fixation (i.e. are outside the "no selection zone"). In other words, the true distribution at least for E coli, is a distribution with a mode at less than zero, a long negative tail, a substantial shoulder either side of the mode that extends into the positive section of the x axis, and includes a tail that extends outside the "no selection zone". And if we read Gerrish and Lenski more closely, we note that this distribution makes very specific assumptions about the environment itself. Sanford seems to ignore the fact that "fitness" as measured the sources he cites in support of his case, is a function of the current environment. And so, in any well-adapted population, the mode will tend to be below zero. Indeed, most lineages derived from Lenski's ancestral colony were a lot fitter than that ancestral colony. In other words, they adapted. But once a population is optimised for an environment, there is very little room for any further improvement, especially in an asexually reproducing species. But change the environment, and the distribution will change as well, obviously. As for epistasis: yes of course, it is in Sanford's index, but turn to his page 110. He regards it as "a sophisticated expression [that] signifies nothing". In other words he ignores it as a factor (except to say that "even if it were valid, it makes the genetic situation worse not better"). And to cap it all, he cites Elena and Lenski (1997) in support of the claim that the "concept is not valid". Elena and Lenski do no such thing. The paper itself is a letter to Nature about specific theory regarding the evolutionary origins of sexual reproduction. They dismiss the theory, not because the "concept [of synergistic epistasis] is not valid" but because both synergistic and antagonistic epistasis occurs in asexual populations:
Because we performed tests for interactions, and not all were independent (using the same nine mutations), we then applied the Bonferroni method to adjust significance levels for the multiplicity of tests. Even with this conservative approach, three synergistic and four antagonistic interactions are significant. Therefore, the mutational deterministic hypothesis seems to fail not because interactions between deleterious mutations are very rare, but rather because synergistic and antagonistic interactions are both common.
Elizabeth Liddle_{June 21, 2011
June
06
Jun
21
21
2011
02:41 PM
2
02
41
PM
PDT}

Elizabeth Liddle:
Then there is the whole issue of epistasis, which Sanford ignores.
According to the index: epistasis: 53, 55, 58, 76, 100, 110, 124, 172, 173, 216, 224. see also Synergistic Epistasis in the Glossary. Ignores? Seriously?Mung_{June 21, 2011
June
06
Jun
21
21
2011
11:46 AM
11
11
46
AM
PDT}

Elizabeth, I am looking at Fig. 3d. The x (horizontal) axis is labeled mutation effect. Sanford has plotted a curve of decreasing frequency (the y axis) as the mutation effect becomes more positive. Yet you claim that Sanford does not do that. You are mistaken.
he takes Kimura’s theoretical model for deleterious mutations only, shows the “VSDM” zone (i.e. those theoretically deleterous mutations that are “effectively” neutral, and then jams next to the zero line on the positive side, the population of observably beneficial mutations (BMs), totally ignoring the nearly neutral zone between zero and BMs(i.e. the VSBM zone).
I see a curve. Don't you see a curve? The frequency decreases as the effect increases. That's what I see. Every one of those graphs has the same label for the x axis (mutation effect). Not only does he show a curve, but the entire curve falls within the "no selection zone"! So you are doubly mistaken. The text for 3d even reads: "it becomes obvious that essentially all beneficial mutations will fall within Kimura's 'no-selection zone'." If you are trusting your memory, let me suggest that you actually refer to the book instead.Mung_{June 21, 2011
June
06
Jun
21
21
2011
11:39 AM
11
11
39
AM
PDT}

That's fine, ba77: I disagree, but it's a fair point. It's just not relevant to Sanford's Figure 3d which is what we were talking about. Selection coefficients (the X axis on the plot) are relative not absolute and they are relative to the current environment. If you want to discuss the issue as to whether a mutation can be beneficial without "breaking" an existing useful function, that's a quite different issue. There are quite a few mechanisms (and I gave several earlier) by which this can occur. One is where genes are duplicated; one is where functions are duplicated; one is where the same function is expressed under additional conditions. There are many others.Elizabeth Liddle_{June 21, 2011
June
06
Jun
21
21
2011
05:39 AM
5
05
39
AM
PDT}

No Elizabeth, it is just plain crazy to say 'breaking stuff will lead to increased functional complexity'. I don't care how you dress it up Elizabeth, to presuppose you can produce functional complexity, which far, far, surpasses anything man has ever done in his most advanced machines, by a process that consistently 'breaks stuff', is to mock logic, defy rationality, and commit yourself firmly to a field of completely unsubstantiated pseudo-sciencebornagain77_{June 21, 2011
June
06
Jun
21
21
2011
03:52 AM
3
03
52
AM
PDT}

But this is a quite different argument, ba77. It is not the argument that Sanford is making in his distribution plots. And is actually irrelevant to them, because "beneficial" in population genetics only means: increases net fitness in the current environment. It doesn't refer to something that would result in net loss of fitness in some other environment. Yes, bridges are often burned (though not always). You evolve a wing, you lose your arms. The birds don't seem to miss them :)Elizabeth Liddle_{June 21, 2011
June
06
Jun
21
21
2011
03:23 AM
3
03
23
AM
PDT}

Elizabeth Liddle, as has been pointed out to you repeatedly, the Lenski 'beneficial mutations' are shown by Behe to in fact be 'loss of function' mutations that are detrimental on a molecular scale. i.e. They are only beneficial in the extremely narrow sense of 'burning a bridge',, Thus it is only by 'craziness' that you can possibly construe these as being positive evidence for neo-Darwinian evolution; Michael Behe’s Quarterly Review of Biology Paper Critiques Richard Lenski’s E. Coli Evolution Experiments – December 2010 Excerpt: After reviewing the results of Lenski’s research, Behe concludes that the observed adaptive mutations all entail either loss or modification–but not gain–of Functional Coding ElemenTs (FCTs)bornagain77_{June 21, 2011
June
06
Jun
21
21
2011
03:10 AM
3
03
10
AM
PDT}

Hi, Bruce! I assume you are looking at Figure 3d right? What he has done is to take Kimura's theoretical monotonic (exponential, I think) curve on the left of zero, and then, to the right of zero, still within the "no selection zone" added a tiny (but, he says drawn "relatively large") triangle representing the 1-in-a-million "beneficial mutations" estimated by Lenski. But these "1-in-a-million" BMs estimated by Lenski (with very important caveats - he is talking about an already adapted population) are those that are selected, they are not those in the "no selection zone. So Sanford's tiny triangle needs to be moved along to the right outside the shaded zone. Then, if he joined the mode at zero to the triangle in its new position he'd have something a bit more sensible, but it would, unfortunately, undermine one of his points, i.e. that even the beneficials are unselectable. It's the selectable beneficials that Lenski counted. But in any case, we have actual data from Elena et al, demonstrating that what I have described above is reflected in actual data. And the point about the "no selection" zone is that drift still occurs. If only VSDMs were in the "no selection zone" Sanford might conceivably have a point, but because he leaves out the VSBMs, which are also subject to drift, that point is somewhat cancelled. But the most important thing that Sanford ignores is that the selection coefficient of an allele is itself a function of the environment, and the environment is not only constantly changing (and if it changes, the number of BMs will go up, as I pointed out above), but the evolving population itself is part of that environment. And that is what the very interesting Gerrish and Lenski paper is about. Because they froze the ancestral population, they can identify whether any new mutation in one lineage is beneficial relative to the ancestral population. And what they find is quite a few beneficials. However, when these beneficials have to compete with other individuals bearing different beneficials, then they may be relatively neutral,or even deleterious! This isn't epistasis of course, but what they term "clonal interference". I have the Gerrish and Lenski paper(also the Elena paper), btw, so if you want to contact me (best way is by PM at Talk Rational - I'm "Febble") I could send it to you. But do remember that Kimura's exponential is a mathematical convenience, to give a monotonic function at values less than 0. A distribution across zero wouldn't peak at infinity! Also, the mode is probably not at zero (probably below, for a well-adapted population).Elizabeth Liddle_{June 21, 2011
June
06
Jun
21
21
2011
12:20 AM
12
12
20
AM
PDT}

Elizabeth Liddle:
I’ve read Sanford’s book in great detail...
In that case, how can you say this?
Then there is the whole issue of epistasis, which Sanford ignores.

Why should mutations increase exponentially (Kumura’s assumption) in frequency as deleteriousness goes from Deleterious to Slightly Deleterious to Very Slightly Deleterious as they approach zero, and then jump straight to “Beneficial” without passing through Very Slightly Beneficial, to Slightly Beneficial on the way?
When I look at the graph I see a curve. So it looks to me like Sanford is doing what you claim he does not do, which is show a change in the numbers through various stages of very slightly beneficial to beneficial. So I think you're misrepresenting Sanford. There;s a line in the middle. Can we assume that's neutral? There's a curve on the left of he line, which is deleterious, and a curve on the right of the line, which is beneficial. Do you not see a curve on the right as well?Mung_{June 20, 2011
June
06
Jun
20
20
2011
08:38 PM
8
08
38
PM
PDT}

I'm touched, Bruce :) OK, point by point:
I concluded that you were defending a position rather than genuinely seeking truth mainly on the basis of your characterization of Sanford’s construction of the distributions illustrated in 3a through 3d. It seems to me quite obvious why he has constructed them the way he has, particularly 3d, and my reading is that he has justified his conclusions quite thoroughly. What he has done in 3d, as I see it, is to take Kimura’s graph, which is essentially the distribution of negative mutations, both selectable and “near neutral”, and add the beneficial mutations to it. And he assumes the exact same shape of distribution. However, as he states quite clearly, since the ratio of detrimental to beneficial mutations is at least 10,000 to one, an assertion he backs up with several different references and arguments, the beneficial mutations will be invisible on the graph at the scale at which it is drawn. So he includes a tiny triangle on the right side to represent the distribution of beneficial mutations. So your statement in #65, “Obviously he assumes it doesn’t exist, but he gives absolutely no reason for this assumption, and it makes no sense,” seems completely unjustified to me.
OK, well let me try to explain exactly what I meant. First of all, the Kimura paper from which Sanford takes the exponential plot is a completely theoretical paper (and none the worse for that) about drift in deleterious mutations including mutations that have selection coefficients at or near zero. It is not a data paper. It makes the math much easier to stop at zero, because that way you can model the distribution below zero as an monotonic function, which he does. Then, Sanford quotes Gerrish and Lenski, in another partly theoretical paper, but one that is tested on actual data. Their paper is not about deleterious mutations but about beneficial mutations, specifically, beneficial mutations in asexually reproducing populations, in which there is no cross-over. In such populations, beneficial mutations will tend to “compete”, and leading to a “law of diminishing returns” whereby the greater the beneficial mutation rate, the smaller the probability that a new beneficial mutation will go to fixation in the population. Interestingly, Gerrish and Lenski assume a monotonic (exponential, in fact) distribution for beneficial mutations for exactly the same reasons as Kimura assumes one for deleterious ones. In support of this, they cite Elena et al (1998) who explicitly studied the distribution of selection coefficients in E coli, and found a negatively skewed distribution of selection coefficients with a long negative tail (severely deleterious mutations) and a mode that was less than 1 (i.e. deleterious); however there is no “cliff edge” at one – there is monotonic, exponential-type decrease, albeit slightly steeper, but which is substantially non-zero at values above 1. So Sanford seems to have regarded Kimura’s theoretical exponential <1 plot as indicating that the frequencies above 1 (not even shown) are neglible, and then grafted on to it a number derived from Gerrish and Lenski, without apparently having read the the Gerrish and Lenski paper itself, and having ignored both the cited Elena datapaper, and, indeed, common sense!
“As for the mutagenic experiments, they are irrelevant. Blasting fruitflies with radiation isn’t going to produce more ‘slightly beneficial’ or ‘slightly deleterious’ mutations – it’s going to produce grossly deleterious mutations.”
How do you know this? It seems to me that increasing the rate of mutation with mutagenic agents should produce all manner of mutations–severe, near neutral, and everything in between. Do you have a reference that supports your claim?
Not to hand, but I guess I could dig some out. But think about it logically: if you increase the rate of mutation dramatically above normal levels, you will certainly induce a much larger rate of severely deleterious mutations. And if your population now bears large numbers of severely deleterious mutations, the odd sweetener in the form of a slightly beneficial mutation isn’t going to do it much good. After all, to take an extreme case, if a mutation renders you infertile, another mutation that makes you slightly more attractive to a mate isn’t going to increase the number of offspring you leave! This is why epistasis is so important, is what the Gerrish and Lenski paper is all about, and which Sanford completely ignores, except when he tackles beneficial mutations (see below) and gets it wrong!.
Regarding the idea that a shift in the environmental conditions can shift the graph, my understanding, based on Behe’s work and other sources I have read (including Sanford), all the actual experimental and observational evidence supports the conclusion that an increase in fitness in a species to adapt to an environmental stress is universally accomplished by breaking existing genes in some way, resulting in a net loss of biological function. This is true for Lenski’s bacteria, insects adapting to pesticides, bacteria adapting to antibiotics, humans adapting to Malaria (sickle cell anemia), and the malarial parasite adapting to anti-malarial drugs. So while it is true that in a stressful environment a detrimental mutation can increase fitness for that particular environment, the evidence suggests that it is virtually always by decreasing the amount of functional information in the genome, and thus in the long run contributing to genetic entropy. Note also that these examples do not support the contention that the entire graph has shifted to the left, but rather that certain few (very few) mutations have crossed over the line from detrimental to beneficial.
Well, what I would say is that Sanford fails to recognise that “fitness” is always computed relative to the current environment. It makes no sense in population genetics terms to say that a mutation increases fitness but is “really” detrimental. If it increases fitness it increases fitness. It may be fair to say that it increases net fitness according to a law of diminishing returns (and this is what the Gerrish and Lenski paper was about) – a beneficial mutation may replace a mutation that did something else quite useful but was not quite as beneficial. But this isn’t always even the case. In one of the recent Lenski lab papers reviewed here, one mutation was not beneficial at all until another came along, and then it was (so epistasis can be positive as well as negative). And not all mutations involve the loss of something – a mutation can happen in a duplicate gene, or a duplicate function. Or it can simple usefully extend the range of conditions under which a gene is expressed. Or, in the case of sickle cell, a mutation can be helpful if heterozygotic but harmful if homozygotic. And this is simply the special case of what are probably numerous gene-gene interactions – it’s the cocktail you inherit that decrees your fitness, rather than any one gene. A gene may be beneficial in genotype, but deleterious in another. So not only is the selection coefficient a function of the current external environment, it’s a function of the genetic environment too. Epistasis is complicated!
So, Lizzie, I take back my accusation of your being in denial. However, I don’t see that you have really understood Sanford’s argument, as I have explained above. It’s refreshing discussing these topics with you, really. Bruce
And with you :) I had to rush the last part of my post, so I hope it makes sense. Feel free to probe if it doesn’t :)Elizabeth Liddle_{June 20, 2011
June
06
Jun
20
20
2011
10:09 AM
10
10
09
AM
PDT}

Ah, Lizzie, you have won me back, you have. I am so used to people in these blogs defending a position instead of really considering the points that have been made, that I find it quite refreshing that you don't seem to be in that mold. I have been in your position (that is, responding to several adversaries with no allies) a number of times when the subject matter is theology rather than materialism and/or Darwinism, as I am a theist but not a Christian, and my views are substantially at odds with most if not all of the Christian theists who post and comment here. It is frustrating when one makes a very good point and it is simply not seen because people's commitment to their world view is so strong that they cannot admit any possible valid challenge to it. I concluded that you were defending a position rather than genuinely seeking truth mainly on the basis of your characterization of Sanford's construction of the distributions illustrated in 3a through 3d. It seems to me quite obvious why he has constructed them the way he has, particularly 3d, and my reading is that he has justified his conclusions quite thoroughly. What he has done in 3d, as I see it, is to take Kimura's graph, which is essentially the distribution of negative mutations, both selectable and "near neutral", and add the beneficial mutations to it. And he assumes the exact same shape of distribution. However, as he states quite clearly, since the ratio of detrimental to beneficial mutations is at least 10,000 to one, an assertion he backs up with several different references and arguments, the beneficial mutations will be invisible on the graph at the scale at which it is drawn. So he includes a tiny triangle on the right side to represent the distribution of beneficial mutations. So your statement in #65, "Obviously he assumes it doesn’t exist, but he gives absolutely no reason for this assumption, and it makes no sense," seems completely unjustified to me. "As for the mutagenic experiments, they are irrelevant. Blasting fruitflies with radiation isn’t going to produce more 'slightly beneficial' or 'slightly deleterious' mutations – it’s going to produce grossly deleterious mutations." How do you know this? It seems to me that increasing the rate of mutation with mutagenic agents should produce all manner of mutations--severe, near neutral, and everything in between. Do you have a reference that supports your claim? Regarding the idea that a shift in the environmental conditions can shift the graph, my understanding, based on Behe's work and other sources I have read (including Sanford), all the actual experimental and observational evidence supports the conclusion that an increase in fitness in a species to adapt to an environmental stress is universally accomplished by breaking existing genes in some way, resulting in a net loss of biological function. This is true for Lenski's bacteria, insects adapting to pesticides, bacteria adapting to antibiotics, humans adapting to Malaria (sickle cell anemia), and the malarial parasite adapting to anti-malarial drugs. So while it is true that in a stressful environment a detrimental mutation can increase fitness for that particular environment, the evidence suggests that it is virtually always by decreasing the amount of functional information in the genome, and thus in the long run contributing to genetic entropy. Note also that these examples do not support the contention that the entire graph has shifted to the left, but rather that certain few (very few) mutations have crossed over the line from detrimental to beneficial. So, Lizzie, I take back my accusation of your being in denial. However, I don't see that you have really understood Sanford's argument, as I have explained above. It's refreshing discussing these topics with you, really. BruceBruce David_{June 19, 2011
June
06
Jun
19
19
2011
11:08 PM
11
11
08
PM
PDT}

Elizabeth, perhaps actual evidence for your position would help???bornagain77_{June 19, 2011
June
06
Jun
19
19
2011
01:50 PM
1
01
50
PM
PDT}

Well, no, Bruce, I don't have a "similar opinion of you". I tend to assume that if something that seems obvious to me is not obvious to someone else, it's either because they are in possession of some information that I don't have, or I am in possession of some information they don't have. Occasionally, I come to the conclusion that they really have some psychological block or something, but it's rare. Which is why it puzzles me when people seem so readily to jump to that conclusion. I've read Sanford's book in great detail; I've read the papers he cites. It's my considered view that he has misunderstood those papers and, in addition, made a number of flawed arguments, even given what I consider are his flawed premises. You may well disagree, and of course I won't continue to press the point with you if you don't want to discuss it. But it does sadden me, I admit, when someone jumps to the conclusion that I am in "denial" rather than at least investigate the alternative, which is that I might have good reasons for my views (even if those reasons don't turn out to be good enough). But, FWIW, I have not jumped to that conclusion wrt to you. Which is why I find it somewhat frustrating that rather than explain to me why you think my point about Kimura's graph is wrong, and why you think that Lenski's likelihood estimate is "speculation" (yet Kimura's entirely theoretical model is apparently not), you simply assume I am in "denial" and leave it at that. :( Still, I guess it was nice to get this far :) Cheers LizzieElizabeth Liddle_{June 19, 2011
June
06
Jun
19
19
2011
01:31 PM
1
01
31
PM
PDT}

Ok, Lizzie, I'm not going to argue with you any more. It is my opinion that you are unwilling to see the obvious, but clearly nothing I can say is going to convince you of that, and I'm sure you have a similar opinion of me, so I guess we'll just have to agree to disagree. Cheers, BruceBruce David_{June 19, 2011
June
06
Jun
19
19
2011
12:23 PM
12
12
23
PM
PDT}

Bruce, the word "likely" does not indicate that a finding is "speculation" rather than "data". A likelihood estimate is often derived from data, and also often derived from theoretical models that may later be tested against data. The series of Figures 3a to 3d are entirely theoretical; the ones based on Kimura's paper are based on Kimura's theoretical model. Lenski's "likehoood" is based on another theoretical model. But the biggest error is Sanford's 3d which is neither Kimura's nor Lenski's but something he has cobbled together from both, and based, as I tried to explain, IMO, on a misunderstanding of both. As for "assuming" vs "deducing": he does note "deduce" the cliff edge. He simply assumes that it must be there. For some bizarre reason he takes Kimura's theoretical model for deleterious mutations only, shows the "VSDM" zone (i.e. those theoretically deleterous mutations that are "effectively" neutral, and then jams next to the zero line on the positive side, the population of observably beneficial mutations (BMs), totally ignoring the nearly neutral zone between zero and BMs(i.e. the VSBM zone). Obviously he assumes it doesn't exist, but he gives absolutely no reason for this assumption, and it makes no sense. Why should mutations increase exponentially (Kumura's assumption) in frequency as deleteriousness goes from Deleterious to Slightly Deleterious to Very Slightly Deleterious as they approache zero, and then jump straight to "Beneficial" without passing through Very Slightly Beneficial, to Slightly Beneficial on the way? It makes no sense at all, and there is neither data nor theory to support it. At least none that I know of and none that Sanford provides. As for the mutagenic experiments, they are irrelevant. Blasting fruitflies with radiation isn't going to produce more "slightly beneficial" or "slightly deleterious" mutations - it's going to produce grossly deleterious mutations. As for the "Loss of function" argument; it is of course true that sometimes a new function is gained at the expense of an old one, and if the new one is more beneficial than the old one in the current environment,then the new one will tend to propagate through the population. However, this is not always the case; sometimes the new function is an additional use for an old function (gene is expressed under a new set of conditions); sometimesthe new function is a use for a gene that has been inactive or irrelevant for a while; sometimes the new function replaces a function that is duplicated elsewhere (may even be a duplicate gene, but not always; most functions are polygeneic). Then there is the whole issue of epistasis, which Sanford ignores. And of course there is the absolutely crucial point that I made in my earlier post that the mode of the distribution itself will move depending on the environment! So far from being "in denial" (I do wish people wouldn't assume that if someone disagrees with them, they are "in denial" :)) as I see it neither Sanford's argument nor his evidence support his overall case at all. He has a couple of good points, the best of which that it is true that in a small effective population, genetic entropy becomes a problem. That is why inbreeding is problematic, and why species may start to become endangered even when there are many hundreds of individuals still in existence. It is also true that in a relatively benign environment, mutations that are harmless in that environment may build up, and prove deleterious if the environment suddenly becomes much harsher. That is probably an important factor in extinctions. So both those points are good. But Sanford draws inferences about "Genetic Entropy" that go well beyond those conditions, and these are simply not justified either by theory or data.Elizabeth Liddle_{June 19, 2011
June
06
Jun
19
19
2011
10:20 AM
10
10
20
AM
PDT}

Lizzie: In #54 you state, "Firstly , it is true that he assumes a cliff edge (See Figure 9)." In figure 3d (p. 32 of the paperback edition), he adds the beneficial mutations to Kimura's distribution, although he can't draw it to scale because the detrimental mutations outnumber the beneficial ones by between 10,000 and 1,000,000 to one. He doesn't "assume" a "cliff edge", he deduces it from the magnitude of that difference, which he gets straight from the literature. Regarding #55, your quote from Gerrish and Lenski contains the telling phrase, "It is likely", which signals that they have left the arena of hard data and entered the realm of speculation--speculation, I submit, that is heavily influenced by an a priori commitment to a naturalistic, Darwinian explanation. Their belief on this point is undermined by the results from Lenski's own experiments on bacteria plus the results of mutagenic experiments on fruit flies and plants, namely that all "beneficial" mutations in these experiments came as a result of genome degradation--loss of information and loss of biological function. So let's take the data from the OP--each new generation of humans contains on average 60 new mutations. Now, we can assume that those that get passed on to future generations are "near neutral", since those that can be "seen" by selection will be selected out. However, since the ratio of detrimental to beneficial mutations, even "near neutral" ones, is at least 10,000 to one, we can assume that all 60 are detrimental. So if a generation is 25 years, then in one million years, there will be 250,000 X 60 = 15 million near neutral BUT deleterious mutations being carried in the genomes of each human being. Now we don't know enough yet about how the genome works to be able to calculate with certainty what the effect of such a load would be, so Sanford invokes an analogy: one "mutation" in, say, the Encyclopedia Britannica will degrade one word or duplicate a paragraph, but will have minimal effect on the work's comprehensibility. However, there will certainly come a point where an accumulation of such errors will seriously impair its usefulness, and there will also come a point where it has become incomprehensible. Now I know that an analogy is not a proof, and you can continue to believe that this relentless accumulation of mutations will not have lethal consequences in the long run, but Lizzie, if you do, I will have to agree with Bornagain (although I don't agree with his style) that you are in denial.Bruce David_{June 18, 2011
June
06
Jun
18
18
2011
10:28 PM
10
10
28
PM
PDT}

What do you mean by “vertical evolution”, bornagain77?
That's when you turn the petri dish on it's side.Mung_{June 17, 2011
June
06
Jun
17
17
2011
12:49 PM
12
12
49
PM
PDT}

Dang it Elizabeth, man I want to see some fireworks: Lenski DID NOT pass the fitness test against the parent strain in native environment!!! Is Antibiotic Resistance evidence for evolution? - 'The Fitness Test' - video http://www.metacafe.com/watch/3995248 Thank Goodness the NCSE Is Wrong: Fitness Costs Are Important to Evolutionary Microbiology Excerpt: it (an antibiotic resistant bacterium) reproduces slower than it did before it was changed. This effect is widely recognized, and is called the fitness cost of antibiotic resistance. It is the existence of these costs and other examples of the limits of evolution that call into question the neo-Darwinian story of macroevolution. http://www.evolutionnews.org/2010/03/thank_goodness_the_ncse_is_wro.html Michael Behe's Quarterly Review of Biology Paper Critiques Richard Lenski's E. Coli Evolution Experiments - December 2010 Excerpt: After reviewing the results of Lenski's research, Behe concludes that the observed adaptive mutations all entail either loss or modification--but not gain--of Functional Coding ElemenTs (FCTs) http://www.evolutionnews.org/2010/12/michael_behes_quarterly_review041221.html Lenski's e-coli - Analysis of Genetic Entropy Excerpt: Mutants of E. coli obtained after 20,000 generations at 37°C were less “fit” than the wild-type strain when cultivated at either 20°C or 42°C. Other E. coli mutants obtained after 20,000 generations in medium where glucose was their sole catabolite tended to lose the ability to catabolize other carbohydrates. Such a reduction can be beneficially selected only as long as the organism remains in that constant environment. Ultimately, the genetic effect of these mutations is a loss of a function useful for one type of environment as a trade-off for adaptation to a different environment. http://www.answersingenesis.org/articles/aid/v4/n1/beneficial-mutations-in-bacteria Elizabeth, that is a big fat dud as far a evolutionary fireworks are concerned. Think about it Elizabeth, the programming in DNA is vastly superior to anything man has yet programmed!!! Systems biology: Untangling the protein web - July 2009 Excerpt: Vidal thinks that technological improvements — especially in nanotechnology, to generate more data, and microscopy, to explore interaction inside cells, along with increased computer power — are required to push systems biology forward. "Combine all this and you can start to think that maybe some of the information flow can be captured," he says. But when it comes to figuring out the best way to explore information flow in cells, Tyers jokes that it is like comparing different degrees of infinity. "The interesting point coming out of all these studies is how complex these systems are — the different feedback loops and how they cross-regulate each other and adapt to perturbations are only just becoming apparent," he says. "The simple pathway models are a gross oversimplification of what is actually happening." http://www.nature.com/nature/journal/v460/n7253/full/460415a.html ,,, yet here you sit Elizabeth with 5 pitiful excuses for beneficial (loss of function) mutations after the equivalent of 2,000,000 years of human evolution!!!! This is not impressive Elizabeth to put it very mildly!!!!bornagain77_{June 17, 2011
June
06
Jun
17
17
2011
08:58 AM
8
08
58
AM
PDT}

What do you mean by "vertical evolution", bornagain77? And at 59: the 5 "beneficial" mutations found by Lenski et al were not VSBMs - if they had been, they wouldn't have been found! They were plain old BMs. And if, by "vertical" evolution, you mean population change in fitness down a single lineage, then Lenski's E-coli are an example of "vertical" evolution. Brand new mutations (not pre-existing) resulted in increased fitness.Elizabeth Liddle_{June 17, 2011
June
06
Jun
17
17
2011
08:23 AM
8
08
23
AM
PDT}

Basically Elizabeth, I just want to see some vertical evolution instead of talk so that you may even have a pretense of being scientific!!!bornagain77_{June 17, 2011
June
06
Jun
17
17
2011
08:12 AM
8
08
12
AM
PDT}

Elizabeth, just one little problem in your trying to maneuver the data to your desired conclusion of VSBMs, why in the world were only 5 beneficial (loss of function) mutations found by Lenski, after billions upon billions of mutations after 50,000 generations, to use in his most recent experiment???. If the plethora of VSBM were truly as great as you imagine them to be, with a simple change in environment, should not you have some ACTUAL evidence for your desired conclusion, instead of just high talk and shuffling of boundaries on paper???bornagain77_{June 17, 2011
June
06
Jun
17
17
2011
08:09 AM
8
08
09
AM
PDT}

Thanks Mung. Made my day :)Elizabeth Liddle_{June 17, 2011
June
06
Jun
17
17
2011
08:01 AM
8
08
01
AM
PDT}

Well mung, once again we disagree.bornagain77_{June 17, 2011
June
06
Jun
17
17
2011
08:00 AM
8
08
00
AM
PDT}

Hi Elizabeth. I really liked your post at 54. I think it's well-reasoned and that you may well have some valid points. BA77:
Mung, If it is truly a random mutation within a functional sequence, such as a typo in a sentence, then you are clearly moving away from optimal functionality to sub-optimal functionality and are thus ‘slightly detrimental’. If it is ‘sub-functional or non-functional’ energetic burden will increase and decrease fitness thus favoring removal by further mutation/selection.
Well, BA, One of the things I cared for the least in Sanford's book was his attempt to equate mutations with changes to English text, so I don't find your use of the same problematical analogy compelling either. If a random mutation isn't the same as a typo then what are you left with? With most typos, I can still figure out what the intent was, iow, I can tell it's a typo. I've already dealt with the fallacy of the "energetic burden" argument.Mung_{June 17, 2011
June
06
Jun
17
17
2011
07:53 AM
7
07
53
AM
PDT}

On closer reading, it may be that Gerrish and Lenski's estimate is based on all mutations, not the just the tails. However, my point remains: that you cannot extrapolate from the total numbers north and south of neutral to the numbers only a little bit north and south of neutral, and, in a skewed distribution, you must not. With regard to my point about a changing environment - from Gerrish and Lenski:
A second assumption of our analyses is that neither the beneficial mutation rate nor the distribution of selection coefficients changes over time. But in a constant environment, a population becomes better adapted with time, leaving progressively less room for further improvement. It is likely that a welladapted population has (i) a lower overall rate of beneficial mutation, (ii) a smaller average effect of beneficial mutations, or both. Consequently, = (w) may be a decreasing function of fitness, whereas = (w) may increase with fitness. These parameters are therefore constant only when w is constant. This condition may be met in an environment that changes just fast enough to counter adaptation of a population.
http://myxo.css.msu.edu/lenski/pdf/1998,%20Genetica,%20Gerrish%20&%20Lenski.pdfElizabeth Liddle_{June 17, 2011
June
06
Jun
17
17
2011
12:45 AM
12
12
45
AM
PDT}

Bruce: You write:
Elizabeth, you state: “[Sanford] seems to think that the closer you get, from the deleterious end, to neutrality, the more mutations you find (‘VSDMs’); then, without any justification, whether by argument or evidence, he assumes a cliff edge with virtually no mutations that are ‘VSBMs’.”
This is simply not true.
Firstly , it is true that he assumes a cliff edge (See Figure 9). Secondly, Gerrish and Lenski(1998) are talking about the ratio of observably beneficial mutations to observably deleterious, not "VSBMs" to "VSDMs", which are, by definition, not observable, and for a well-adapted population (at least of E-coli) their estimate may well be correct. We would certainly expect the distribution of all mutations to have a very long negative (deleterious) tail, and possibly not much of a beneficial tail at all. However, the vast majority of mutations will be near-neutral, i.e. not detectably beneficial or deleterious and thus "invisible to selection". Sanford agrees, indeed it is one of the major themes of his book. But he extrapolates from the perfectly good evidence that DMs vastly outnumber BMs (which he doesn't worry about because, as DMs are, by definition "visible to selection" they won't accumulate) to assuming that the same ratio applies to VSDMs and VSBMs. This assumption is completely unjustified, and there is no evidence for it, nor any theoretical reason to believe it (quite the reverse). In a well-adapted population, of course, it is true that there will be more VSDMs than VSBMs because the mode of the distribution will have moved towards the optimum. However, change the environment, and your mode will be shunted to the left, whereupon selection kicks in (because now some VSDMs become SDMs and SDMs become DMs, and similarly, what were only VSBMs or even VSDMs now become SBMs or BMs) and the mode tracks rightward again. And that is Sanford's other error - to forget that whether a mutation is "deleterious" or "beneficial" is not an absolute attribute of the mutation, but is a function of the mutation within its environment. And then there is epistasis, but maybe we'll leave that for another time :) Cheers LizzieElizabeth Liddle_{June 17, 2011
June
06
Jun
17
17
2011
12:13 AM
12
12
13
AM
PDT}

1 2 3 Next

You must be logged in to post a comment.

Leave a Reply