Behe’s “Multiple mutations needed for E. coli”

“But if more than one is needed, the probability of getting all the right ones grows exponentially worse. “If two mutations have to occur before there is a net beneficial effect — if an intermediate state is harmful, or less fit than the starting state — then there is already a big evolutionary problem.””

You are correct that if the intermediate state (after one random mutation) is harmful then the probability that a second muatation will correct it (to produce a more fit individual) is exponentially worse.

However, you don’t mention the converse logic. That is, if the first mutation produces a slightly more fit individual, then the probability is exponentially better for that second mutation to make an even more fit individual.

There is absolutely no evidence to support your assumption that intermediate species are inherently less fit than previous generations. Probably theory states that generation of a less fit individual in this first random mutation is just as likely as generation of a more fit individual.

broadbill

You obviously haven’t read “The Edge of Evolution” for if you had you’d know there is indeed abundant evidence to presume that fitness is compromised in these situations. It’s called “trench warfare”. I’ll take you off moderation after you’ve demonstrated a willingness to read something before you comment on it.

Is this Lenski the guy who promised, a year or more ago, that he had observerd something truly extraordinary in his lab that was going to provide overwhelming proof of Darwinian evolution? Citrate digestion is a real yawner. If that’s all it is, it’s really pathetic and is really a world class example of the lack of any compelling evidence for RM+NS as a mechanism capable of driving creative evolution.

Suppose that the probability of getting the mutations necessary to utilize citrate is 1/30000 generations. Now supposing that instead of just subjecting the cells to an inordinate amount of citrate, what if the environment contained an inordinate amount of 10 different compounds the cells couldn’t utilize, where the probability of the mutations necessary for any one of them averaged the same as for citrate - 1/30000. Then we could expect the time it took for the cells to be able to utilize any one of them would be 1/10 the time it took for the cells to be able to utilize just citrate, so in other words, a couple of years or so.

“No population evolved the capacity to exploit citrate for >30,000 generations, although each population tested billions of mutations” [from Lenski paper abstract]
Who knows how many abilities the cells generated after billions of mutations that were not detected because the experiment wasn’t set up to detect them.

Junkyard Tornado wrote:

“Who knows how many abilities the cells generated after billions of mutations that were not detected because the experiment wasn’t set up to detect them.”

Possibly. But not likely. Remember, the e-coli already had the ability to use the citrate. They just didn’t have a way to efficiently get it into the cell. In other words, if we look at the overall cellular machinery needed for this new function, we’ve already spotted them the ball on the two-yard line and it took 30,000 generations to even get across the goal line. Then millions of bugs were encouraged — all but begged — by the unusual conditions of the experiment to see if they could make it the last tiny step.

This doesn’t mean that novel results aren’t possible. They are, as Behe discusses in his book with malarial resistance. But it shows yet again just how impotent RM+NS is.

broadbill:

the question is quite simple: if E. coli becomes able to feed on citrate after, say, 2 or 3 mutations, that means that until all of them have happened, E. coli can’t feed on citrate. Otherwise we would observe a strain which can moderately feed on citrate, then another one which is better suited to do that, and so on, each one separated from the previous one by a single mutation.

I don’t think that’s what has been observed here. At present we don’t know how many mutations were necessary before the citrate could enter the cell, but if the hypothesis that at least two were necessary is right, that means that the first mutation still had no effect on permeability to citrate.

I agree with you that the first mutation needs not be harmful: it could just have been neutral, as many mutations probably are. But the point is that it was not useful, for itself. Indeed, if it had been useful, it would have been quickly selected, fixed and expanded, and the second mutation would have quickly followed, instead of having to wait for such a long time.

The fact is: a harmful mutation is probably selected against, and has really few chances of surviving long enough to “receive” the second necessary mutation, least of all to be fixed and expanded.

A neutral mutation, on the other hand, is not usually selected. So, it stays confined to just a single individual line. A lot of time is needed so that a second coordinated mutation can happen in that same single line, because the probabilities are extremely low. For three coordinated mutations, each of the two intermediate ones unselected, the probabilities begin probably to be too low even for bacteria.

Let’s remember that there is theoretically a way that a neutral selection could be fixed: it’s genetic drift. But that is a false reasoning. Genetic drift, in the measure it can happen, is completely random. Any single mutation could be fixed by genetic drift, and that means that no single mutation has a special probability of being fixed.

So in the end, if we need two mutations to be present together at some time in one bacterial line, and none of them can be selected if it happens alone, then the probability for the combined mutation is the product of the two probabilities, and it becomes increasingly lower for each new necessary independent, unselected mutation. That’s why 2 or 3 is the empirical limit observed by Behe and, very likely, also by Lenski.

The old model of single step pathways, where each mutation is patiently selected, is only a myth. If it were true, we would easily observe it at work, at least in rapidly reproducing organisms as bacteria and protozoa. Instead, it is never observed, there is no detailed example of it in the empirical realm. The reason is simple: no real new function can be built with a non trivial number of single mutations, each of them creating a gradual increase of function. It’s only a myth, which lives only in the fertile minds of darwinists, and not in the real world.

Eric Anderson: “Remember, the e-coli already had the ability to use the citrate. They just didn’t have a way to efficiently get it into the cell”

Was this ability just part of a generic ability to process various food sources. IOW, its not clear to me whether there was a seperate distinct capability internally tailored to utilize citrate specifically.

“Now, wild E. coli already has a number of enzymes that normally use citrate and can digest it (it’s not some exotic chemical the bacterium has never seen before).

My point, if it wasn’t clear, was the following:

The human digestive system can already digest various things like meat, fruit, and vegetables via the same digestive mechanism (I’m assuming). Suppose you were on a desert island where some rare species of plant predominated that was perfectly edible for humans, except for one specific compound it contained which was highly toxic. If the humans there developed resistance to this toxin via mutations, it would be irrelevant to observe, “Well, humans already had the ability to digest this rare species of plant, and only the toxin kept them from doing so.” The ability was generic.

JunkyardTornado, yes that is the statement that I understood to mean that E. coli already had the ability to use citrate.

Reposting from other thread, per Bill’s instruction:

———

I presume this is the same Richard Lenski who was involved with the Avida silly business? One of the stated purposes of Avida was to show “how complex functions can originate by random mutation and natural selection.”

Boy, it seems a lot harder to evolve novel features in real life than it did with that slick computer program!

It will be interesting to watch this further and see what they ultimately determine was the source of this new ability. Based on the track record, I’ve got to believe that Behe’s intuition about the insignificance of the result is likely spot on.

BTW, for those keeping score, Behe is making a real, albeit softly stated, prediction in his Amazon post. We’ll see who ends up being right.

I particularly love the irony here, as Avida — in my view — inadvertently provided support for Behe’s idea of irreducible complexity. Now Lenski (with E. Coli) will likely end up demonstrating empirically what Behe has been arguing in Edge of Evolution.

JunkyardTornado, sorry it looks we were posting at the same time, so I didn’t get your entire thought the first time around.

I understand your point, but I don’t think that is a fair analogy. Behe says that E. coli has enzymes that normally use and can digest citrate. The challenge seemed to be getting enough of it into the cell.

I’m just going off of the brief info we have at this point. It sounds like Lenski is still working to determine what in fact occurred, so we’ll have to wait and see.

My money is with Behe though on this one . . .

“The ability was generic.”

Just realized that was a pun.
That was a real knee-slapper.

The critical ratio of beneficial to harmful mutations is variously reported as lower than 1 in 10,000, or lower than 1 in one million.
This factor dominates all other parameters.

See discussion

P.falciparum - No Black Swan Observed especially bornagain77’s post 82

The primary thing that is crushing to the evolutionary theory is this fact. Of the random mutations that do occur, and have manifested traits in organisms that can be measured, at least 999,999 out of 1,000,000 (99.9999%) of these mutations to the DNA have been found to produce traits in organisms that are harmful and/or fa^tal to the life-form having the mutation! (Sanford; Genetic Entropy page 38)

“I have seen estimates of the incidence of beneficial mutations which range from one in one thousand up to one in one million. The best estimates seem to be one in one million (Gerrish and Lenski, 1998) Since neutral mutations can be inferred to almost never occur in a genome, then the ratio of deleterious to beneficial mutations seems to be one million to one.” (Sanford; Genetic Entropy, page 38: Note: this statement has been revised to reflect the evolutionary belief of some totally neutral mutations of Gerrish and Lenski)

http://myxo.css.msu.edu/lenski…..Lenski.pdf

Even if there were totally neutral mutations, which is highly unlikely given the overwhelming interrelated complexity of the genome, Gerrish and Lenski most likely used a incomplete measure of fitness/information in order to arrive at their one in a million number for beneficial mutations. I maintain that their, one in a million, estimate for beneficial mutations is flawed and that ALL mutations to a genome will be found to be harmful/fatal when using a correct measure of fitness/information. The following articles points out this flaw, in measuring the total fitness/information of a organism, by evolutionary scientists and thus skewing the already crushing, but biased, mutational studies:

http://www.answersingenesis.or…..vation.asp

http://www.answersingenesis.or…..apter5.asp

” Bergman (2004) has studied the topic of beneficial mutations. Among other things, he did a simple literature search via Biological Abstracts and Medline. He found 453,732 “mutation” hits, but among these only 186 mentioned the word “beneficial” (about 4 in 10,000). When those 186 references were reviewed, almost all the presumed “beneficial mutations” were only beneficial in a very narrow sense- but each mutation consistently involved loss of function changes-hence loss of information.”

In fact, from consistent findings such as these, it is increasingly apparent that Genetic Entropy is the overriding foundational rule for all of biological life with no exceptions at all, and that belief in beneficial mutations is nothing more than wishful speculation that has no foundation in science whatsoever:

The foundational rule for biology can be stated like this:

All adaptations away from a parent species for a sub-species, which increase fitness to a particular environment, will always come at a loss of information from the parent species. (Note: At present viruses are excluded from this rule.)

Professional evolutionary biologists are hard-pressed to cite even one clear-cut example of evolution through a beneficial mutation to DNA that would violate the principle of genetic entropy. Although evolutionists try to claim the lactase persistence mutation as a lonely example of a beneficial mutation in humans, lactase persistence is actually a loss of a instruction in the genome to turn the lactase enzyme off, so the mutation clearly does not violate genetic entropy. Yet at the same time, the evidence for the detrimental nature of mutations in humans is clearly overwhelming, for doctors have already cited over 3500 mutational disorders (Dr. Gary Parker).

“It is entirely in line with the al nature of naturally occurring mutations that extensive tests have agreed in showing the vast majority of them to be detrimental to the organisms in its job of surviving and reproducing, just as changes ally introduced into any artificial mechanism are predominantly harmful to its useful operation” H.J. Muller (Received a Nobel Prize for his work on mutations to DNA)

“But there is no evidence that DNA mutations can provide the sorts of variation needed for evolution… There is no evidence for beneficial mutations at the level of macroevolution, but there is also no evidence at the level of what is commonly regarded as microevolution.” Jonathan Wells (PhD. Molecular Biology)

See especially
Respected Cornell geneticist rejects Darwinism in his recent book

Genetic Entropy & the Mystery of the Genome, by John Sanford (October 2005)

The critical ratio of beneficial to harmful mutations is variously reported as lower than 1 in 10,000, or lower than 1 in one million.

Eh? Can you give me the references that say this? The last review I saw was putting them at a few percent. Still rare, but much more common than you’re suggesting.

p38 of Genetic Entropy doesn’t discuss the proportion of advantageous mutations..

Aagh. On the preview, the link was appending the second double-quote to the link.

This is the last review. The preview still looks wrong, but we’ll see.

Let me make sure I get Behe’s logic right:

1)Multiple mutations are needed for evolution, and are so wildly improbable that evolution can’t occur.

2)Lenski observed an evolutionary event in a relatively short amount of time (think about 20 years in the context of the earth’s history) that, he speculated, may have been caused by (wildly improbable) multiple mutations.

3)Multiple mutations are needed for evolution, and are so wildly improbable that evolution can’t occur.

I’m sure you’ll fill me in on what I’m missing here.

#18

“Let me make sure I get Behe’s logic right:”

Indeed this seems your problem

“1)Multiple mutations are needed for evolution, and are so wildly improbable that evolution can’t occur.”

What does it mean evolution for you? If you had read EOE you should know that what is aty stake here is the possibility for any decent macroevolution. And although both Lenski’s data and Behe arguing about malaria resistance are very under that edge we have evidence that those trivial (trench warfare) evolution is indeed extremely rare.

“2)Lenski observed an evolutionary event in a relatively short amount of time (think about 20 years in the context of the earth’s history) that, he speculated, may have been caused by (wildly improbable) multiple mutations.”

See above; what you call “evolutionary event” is a very trivial one; a mere slight modification on a very complex biochemical stuff

“3)Multiple mutations are needed for evolution, and are so wildly improbable that evolution can’t occur.”

Again see above.

“I’m sure you’ll fill me in on what I’m missing here.”

That’s all folks. But if you had read something more about Behe’s arguments before, you hadn’t asked.

@Dave Scott-
I’ll admit I haven’t read Edge of Evolution but I was commenting on the paraphrase of it in this blog entry. I think my criticism still stands. If I need to go back to the book to have my comment was addressed, then the author of this blog entry is incompletely quoting his source (his own book in this instance)

@gpuccio
As Dave Scott has pointed out, I haven’t read Edge of Evolution so I’ll refrain from commenting on that.
However, I don’t think your statement regarding single mutations being a myth is true in the case of citrate utilization. In these E. coli, the citrate utilization machinery was already there, it just needed a way to get into the cell. In this case, a single mutation in the citrate permease gene may very well caused the phenotype. We will have to see future work by this laboratory to see if this is the case.

Yes, several distinct mutations in gene leading to a production of a novel biochemical pathway would indeed be rare, but utilization of an already existing pathway by a single mutation of a rate limiting enzyme isn’t.

You could also have “hijacking” of existing biochemical pathways to use new substrates. The ability of that enzyme to use that new substrate could be caused by a single mutations.

“Multiple mutations are needed for evolution, and are so wildly improbable that evolution can’t occur.”

Nope, that’s not what he’s saying. He specifically talks about an evolutionary change in the development of Malaria’s resistance to drugs in which two ’simultaneous’ changes must have taken place.

To tired to summarize Behe right now. Have you read Edge of Evolution?

These links, DLH referenced, in 14 work and are worth the read:

Observation of evolution in bacteria

http://www.answersingenesis.or.....vation.asp

excerpt:

One strain had a mutation in a gene for the enzyme glycerol kinase which is important in the first step of glycerol breakdown. This mutation reduced the ability of glycerol kinase to be inhibited by fructose-1,6-bisphosphate (FBP). FBP is important in limiting the rate at which glycerol is catabolized. This is important since a side reaction during glycerol breakdown results in the production of a metabolite which is toxic at high concentrations. No gain of information took place as required by evolution, only loss leading to dysregulation of this pathway. In the wild, versus the rather comfy lab environment, this could be extremely detrimental.

Argument: Some mutations are beneficial

http://www.answersingenesis.or.....apter5.asp

excerpt:

In the process of defending mutations as a mechanism for creating new genetic code, they attack a straw-man version of the creationist model, and they have no answer for the creationists’ real scientific objections. Scientific American states this common straw-man position and their answer to it.

10. Mutations are essential to evolution theory, but mutations can only eliminate traits. They cannot produce new features.

On the contrary, biology has catalogued many traits produced by point mutations (changes at precise positions in an organism’s DNA)—bacterial resistance to antibiotics, for example. [SA 82]

This is a serious misstatement of the creationist argument. The issue is not new traits, but new genetic information. In no known case is antibiotic resistance the result of new information. There are several ways that an information loss can confer resistance, as already discussed. We have also pointed out in various ways how new traits, even helpful, adaptive traits, can arise through loss of genetic information (which is to be expected from mutations).

Mutations that arise in the homeobox (Hox) family of development-regulating genes in animals can also have complex effects. Hox genes direct where legs, wings, antennae, and body segments should grow. In fruit flies, for instance, the mutation called Antennapedia causes legs to sprout where antennae should grow. [SA 82]

Once again, there is no new information! Rather, a mutation in the hox gene (see next section) results in already-existing information being switched on in the wrong place.1 The hox gene merely moved legs to the wrong place; it did not produce any of the information that actually constructs the legs, which in ants and bees include a wondrously complex mechanical and hydraulic mechanism that enables these insects to stick to surfaces.2

These abnormal limbs are not functional, but their existence demonstrates that genetic mistakes can produce complex structures, which natural selection can then test for possible uses. [SA 82]

Amazing—natural selection can ‘test for possible uses’ of ‘non-functional’ (i.e., useless!) limbs in the wrong place. Such deformities would be active hindrances to survival.

—

Only one thing I can add to this, is that it is commonly known that the parent strain of bacteria will consistently be more fit for survival than the mutant strain when compared to the parent strain in the original environment.

Is Antibiotic Resistance evidence for evolution?

http://www.godtube.com/view_vi.....e30ff85177

This very simple demonstration of “demonstrable” evolution in native environment has never been shown. YET….

If evolution were actually true you would naturally expect the “random” mutations of bacteria to, every so often, just randomly develop a complexity in the native environment that surpasses the parent strains complexity and thus surpasses the parent strains ability to survive. Yet this has never been observed! Why must evolutionists always allude to some “shady” characteristic that is in Davescots words “a real yawner” when they should have countless examples of evolution of complexity in native environments for bacteria that would be unambiguous in its proof?

The truth is that they will never demonstrate a gain in complexity for any life-form for the “unmatched” integrated complexity of the information in a life form prevents this from happening. As well I point out that this is able to be inferenced from first principles of science whereas evolution must ignore first principles of science.

broadbill

I’ll admit I haven’t read Edge of Evolution

No admission was required. I wrote that it was obvious you hadn’t read it. It was a statement not a question.

In this case, a single mutation in the citrate permease gene may very well caused the phenotype.

Again, if you’d read The Edge of Evolution you would know better than to write that. The spontaneous single point mutation rate of E.coli vs. the size of its genome coupled with the vast number of individuals in each generation assures us that it will test (over and over and over again) all possible single point mutations in each generation. If a single point mutation with as much benefit as being able to utilize 90% of the available food supply while peers without the mutation starve then that mutant strain will take over the population. The very first culture plate would almost certainly produce citrate eaters.

The math is incontrovertable and is well laid out in Behe’s book.

kairos,

My point is that if Behe is arguing that multiple mutations are too rare to allow evolution of the cell, then it is odd that he is using an example of a multiple mutation (maybe) occurring and, however “trivial” you may consider the mechanism, causing a large increase in fitness through selection (i.e. evolution).

Furthermore, the Lenski article also shows through a nice series of experiments that this evolutionary event was preceded by intermediate steps, which Behe also considers costly and extremely rare.

So he more or less seems to be using a clear example of white to argue black.

dmso74

It’s obvious you have not read The Edge of Evolution either. Why do you people insist on criticizing things you know nothing about? Don’t you realize it makes you look ignorant and lazy?

DaveScot,
I read the book about 3 months ago, after following the back-and-forth over the evolution of the vpu protein in HIV. Can you tell me how I am mis-interpreting it? As I understood it, Behe argues that point mutations are not capable of creating the types of changes needed for evolution.. and the low probability of getting multiple changes (either stimultaneously or through a pattern of intermediates with lowered fitness) dramatically lowers the odds of beneficial phenotypic changes occurring.

This paper, however, shows that a) a series of intermediates with slightly lower or identical fitness and b) (maybe) multiple mutations appears to have led to beneficial change that was strongly directionally selected for. How does this not contradict Behe’s argument?

p.s. Behe limits the number of mutational changes that can occur to two.. however, this paper argues that more than one “potentiating” change occurred early (20,000 generations) that allowed the later beneficial mutation to occur.. if the latter was more than a single mutation, then this goes beyond Behe’s “EoE” threshold.

DLH: “The critical ratio of beneficial to harmful mutations is variously reported as lower than 1 in 10,000, or lower than 1 in one million. This factor dominates all other parameters.
…
The primary thing that is crushing to the evolutionary theory is this fact. Of the random mutations that do occur, and have manifested traits in organisms that can be measured, at least 999,999 out of 1,000,000 (99.9999%) of these mutations to the DNA have been found to produce traits in organisms that are harmful and/or fatal to the life-form having the mutation! (Sanford; Genetic Entropy page 38)
…
I maintain that their, one in a million, estimate for beneficial mutations is flawed and that ALL mutations to a genome will be found to be harmful/fatal when using a correct measure of fitness/information.

——————

Thought I’d run a simple test which has no doubt been done before. I have this application I’ve written which is 529,875 bytes long. That’s optimized, and stripped of all debug and symbolic information. (And of course it accesses various 3rd-party DLL’s as well.) I ran this test where I would change a bit of the program at random (chosen by random number generator) and run the program through a battery of tests.

The application enables the design and building of web pages with various novel graphical effects and textures. So, my test was as follows: I would change a bit at random in the executable file (actually a .DLL) and then do the following: Open a preexisting RTF file. Open an owner-drawn menu three times, each time choosing a different graphical theme for the open file. Then I would change the font of the file. Then I opened another dialog that performed a certain function and test that twice. Finally I would open the dialog to build a web page with the new changes, and from within that open up a subdialog for editing images in the document, close that dialog, and finally hit ‘build’ to build the new webpage. Then I would hit ‘view’ to bring up the new web page in a web browser.

I had decided in advance to perform this test 20 times with 20 different one-bit random changes to the executable. The results: 2 program crashes, 1 malfunction, and in the other 17 cases, no discernable effect whatsoever.

Relevant? Irrelevant?

broadbill:

You say:

“However, I don’t think your statement regarding single mutations being a myth is true ”

But where did I say anything like that? What I said is completely different.

I cite here from ny post:

“The old model of single step pathways, where each mutation is patiently selected, is only a myth.”

As you can see. it is not single mutations which are a myth (IMO), but “pathways” where each single step is selected for function gain. Maybe I did not express myself clearly, I apologize.

I have clearly stated that single mutations are perfectly accessible to all living beings, especially bacteria. Indeed, specific single mutations can happen quite often in bacteria. In another thread, I grossly calculated the probability of a specific mutation in the E. coli genome at 1:(3*4.7 million), that is about 1 : 10^7 for each mutational event. That’s not very low, for a common and fast replicating organism as E. coli.

So, if a specific single mutation cean give an indirect benefit, usually slightly modifying an existing function, it can be selected. All the well documented examples we know od RM and NS are of that kind. Most of them imply single mutations, and almost all of them are examples of indirect advantage, derived from partial degradation of an existing function, selected by special aggressive conditions in the environment (like antibiotics).

I have also said explicitly that two coordinated mutations, that is two mutations which have to be simultaneously present before there is a function gain, are another matter: here the two probabilities multiply, and for E. coli the probability of any specific set of two mutations becomes about 1 : 10^14, which is much lower. Still, that is in the range of bacteria in a reasonable time (decades), while not so much in the range, for in stance, of mammals. Indeed, Behe puts more or less there his “edge” for undirected evolution.

But I want to be more generous. I can accept that, very rarely, specific 3 mutation sets can be attained in bacteria, and selected if they confer gain function. Here the probability becomes 1 : 10^21, and we are already in a really problematic order of magnitude, but you know, luck happens. In bacteria or protozoa, at least. it could happen, although very very rarely. It’s not even the case to discuss higher forms of life here.

But that’s all. If we add more necessary mutations to our set, we are out. That would no more be luck. That would have to be design.

So, if we want to obtain by single independent random mutations a specific set of, say, 10 mutations, the probability would become, in E. coli, 1 : 10^70, and we are out of any reasonable model (I know, we are still not at Dembski’s UPB, which is 1 : 10^150, but I really think we are already out, and I mean really out).

Obviously, there is the alternative possibility that the single steps are selected. That would dramatically make everything infinitely simpler (although not necessarily easy). We eould not have anymore to multiply probabilities, because the expansion of each mutation to all the population would make the probability the same for each new mutation (the previous ones having been fixed). In other terms, it could be done, if we really could justify that kind of single step fixation.

But we can’t. That’s what I meant with my argument about the myth. There is no single pathway known of, say, 10 different mutations, which leads to a definite function gain, and where each of the 9 intermediated is gradually fitter. Even more difficult would be to have such a pathway from one function to a completely different one. Even more difficult would be to have a pathway of, say, 50 mutations (or, as we discussed in detail in the previously mentioned thread, of 490 mutations).

That is the myth: that function landscape can be traverse by specific pathways, where you have a “stepping stone” of higher function at each single mutation, or, if we want to be generous, and if we are discussing bacteria, at every 2-3 mutation distance.

That’s not true. That’s the myth I was alluding to. There is no example of that, neither theorical nor empirical. There is no trace of the billions of functional intermediates that such a scenario would imply.

In other words, that scenario is simply false.

JunkyardTornado (#26):

I think it is relevant. It shows that neutral mutations are quite common, and harmful ones common. Which, IMO, is perfectly true also in biological information.

The real problem comes with beneficial mutations, of course…

junkyardtornado

re; no discernable effect whatsoever

Offhand I’d say you aren’t traversing much of the code in the test. There are tools that help you determine how much of the code has actually been traversed. I’m familiar with this one but it may not be applicable in your environment.

http://www.compuware.com/produ.....c.htm#code

Junkyard you stated:

I had decided in advance to perform this test 20 times with 20 different one-bit random changes to the executable. The results: 2 program crashes, 1 malfunction, and in the other 17 cases, no discernable effect whatsoever.

Relevant? Irrelevant?

Better watch out Junkyard, if evolution is as true as evolutionists assure us it is, you will soon surely discover how to allow computer programs to write themselves. Thus putting many well paid software programmers out of jobs. (You might even crash the entire economy) I might add, when you discover the correct evolutionary process for computer program writing, they (the programs) will write themselves with a level of complexity that we will not be able to understand since the complexity in genomes is currently far beyond man’s grasp to understand. (ENCODE; Bill Gates)

gpuccio: As far as the 17 neutral - Of course nothing I think is really neutral. Each of these I would say had some marginal effect on program performance (in terms of speed or space usage), either good or bad. It just wasn’t discernable from the standpoint of fitness (if user perception is the metric gauging fitness.) Furthermore to assume that I as a designer was so infallible that any change to my code was likely to be at least a marginal decrease in performance would not be reality.

I hadn’t decided before the test what would indicate program improvement, so the test was really only detecting neutral or harmful mutations. The test did not indicate the rarity of beneficial mutations.

DaveScot:

As far as the traversal of code, if I were to devote another hour or two to this I should probably do a code coverage analysis using that tool. I know that at least 80% of the source files were being hit.

Undoubtedly some of those neutral changes probably were doing something pretty harmful. But I was able to perform the program’s primary function without incident. Compare that to getting to breeding age without incident. In another thread, scordova talked about how even obviously beneficial mutations may never have a discernable impact. The same can be said for obviously harmful mutations.

Not an exhautive test I’ve done though, in just ninety minutes.

It is very hard to establish neutrality in deeply redundant systems. You can knock out 1 of the 5 navigation systems of a Space Shuttle, but becuase of the quintuple redundancy, there is no obvious effect on the behavior.

Biological systems are deeply redundant, and system behaviors of the redundant system resist easy characterization by selection. See: Airplane Magnetos.

“Previous attempts to work out the minimal genome have relied on deleting individual genes in order to infer which genes are essential for maintaining life,” said Professor Laurence Hurst from the Department of Biology and Biochemistry at the University of Bath.

“This knock out approach misses the fact that there are alternative genetic routes, or pathways, to the production of the same cellular product.

When you knock out one gene, the genome can compensate by using an alternative gene.

But when you repeat the knock out experiment by deleting the alternative, the genome can revert to the original gene instead.

Using the knock-out approach you could infer that both genes are expendable from the genome because there appears to be no deleterious effect in both experiments.”

What is true of knockout approaches is true of deleterious mutations in regions of deep redundancy.

We shouldn’t presume that lack of immediate effects on fitness are necessarily indicative that a mutation is truly neutral.

junkyard

Still doesn’t sound right but without knowing the code I don’t have much to go on. A lot of the DLL could be data where a flipped bit might only cause some subtle error - a misspelled word or a pixel that is out of place.

Try this instead. Randomly flip a bit in the source code instead of the machine code (but make sure it doesn’t land in a comment). I’d bet dollars against donuts in more cases than not you won’t even be able to test the executable because you’ll get a compiler error from the altered source code. One thing you can be sure of is that the compiler will traverse every bit of the source code that isn’t a comment.

DaveScot:
“Still doesn’t sound right but without knowing the code I don’t have much to go on. A lot of the DLL could be data where a flipped bit might only cause some subtle error - a misspelled word or a pixel that is out of place.”

You could say the same about random changes to the genome - 20 changes at random are likely to a subtle errors-a pixel out of place, etc. with no real impact.

If statically allocated strings are being altered I would have reported misspellings. Also if there are statically allocated arrays and the like, its not like there will be strings of 0’s or something in the actual executable file. The data segment will be allocated at program startup. The exe will contain code, resources like dialogs, as well as staticly allocated strings, etc.

Anyone could perform this test. Just bring up some executable in a hex editor (download something from cnet.com) and start making random changes. However, they would have to be actually random.

Will get back to this in a bit, however.

Or if anybody is interested in seeing a demo of my program, that could possibly be arranged as well.

Dave Scot:

I recall you were involved with hardware design or assemblers or something, so you probably have better intuition even than me as to why you can start changing bits around in machine code and be unscathed. I just think in terms like - if a word in hardware is n bits and your instruction set only requires n-m bits, then this means the top n-m bits of any instruction are being ignored. I’m sure there are many other factors as well. On substantive areas, if you allocate a dynamic array of 0xACDF bits (whatever that is in decimal) and a random bits change increases the size to 0xBCDF, then no harm. Also, I’m sure when you compile something, the compiler generates all sorts of boilerplate machine code, generalized for a number of potential scenarios only a handful of which may ever materialize in one particular program. This code could also be mangled and it not make any difference. Scordova mentioned redundancy in the genome, so if there is redundancy there, in the form of junk-dna or whatever obviously that can be mangled and not make a difference.

As far as compilers and high-level languages, I don’t think nature is such that if one tiny thing is out of place it just shuts down and refuses to do anything. It will take what you give it and attempt to do something, which is what a computer processor does, I think.

Don’t know how much light this analyis sheds.

The scenario I tested says whatever it says.

Of interest to topic:

Page3 20=21 Genetic Entropy ; Sanford;

Are there truly neutral nucleotide positions? True neutrality can never actually be demonstrated experimentally (it would require infinite sensitivity). However, for reasons we will get into later, some geneticists have been eager to minimize the functional genome, and wanted to regulate the vast bulk of the genome to “junk DNA”. So mutations in such DNA would be assumed to be entirely neutral. However actual findings relentlessly keep expanding the size of the functional genome, while the presumed “junk DNA” keeps shrinking. In just a few years, many geneticists have shifted from believing that less than 3% of the total genome is functional, to believing that more than 30% is functional - and that fraction is still growing. As the functional genome expands, the likelihood of neutral mutations shrinks. Moreover, there are strong theoretical reasons for believing there is no truly neutral nucleotide position. By its very existence, a nucleotide position takes up space, affects spacing between other sites, and affects such things as regional nucleotide composition, DNA folding and nucleosome binding. If a nucleotide carries absolutely zero information, it is then by definition slightly deleterious - as it slows cell replication and wastes energy. Just as there are really no truly beneficial neutral letters in a encyclopedia, there are probably no truly neutral nucleotide sites in the genome. Therefore there is no way to change any given site, without some biological effect - no matter how subtle. Therefore, while most sites are probably “nearly neutral”, very few, if any, should be absolutely neutral.

Since Dr. Sanford made thus crushing critique in 2005 the genome of humans has now been shown to be virtually 100% severely poly-functional, by ENCODE, with no “junk DNA” regions. Thus this principle is devastating to evolutionary theory (but boy do they a song and dance around it!)

This “complex interwoven (poly-fuctional) network” throughout the entire DNA code makes the human genome severely poly-constrained to random mutations (Sanford; Genetic Entropy, 2005; page 141). This means the DNA code is now much more severely limited in its chance of ever having a hypothetical beneficial mutation since almost the entire DNA code is now proven to be intimately connected to many other parts of the DNA code. Thus even though a random mutation to DNA may be able to change one part of an organism for the better, it is now proven much more likely to harm many other parts of the organism that depend on that one particular part being as it originally was. Since evolution was forced, by the established proof of Mendelian genetics, to no longer view the whole organism as to what natural selection works upon, but to view the whole organism as a multiple independent collection of genes that can be selected or discarded as natural selection sees fit, this “complex interwoven network” finding is extremely bad news, if not absolutely crushing, for the “Junk DNA” population genetics scenario of evolution (modern neo-Darwinian synthesis) developed by Haldane, Fisher and Wright (page 52 and 53: Genetic Entropy: Sanford 2005)!

http://www.genome.gov/25521554

BETHESDA, Md., Wed., June 13, 2007 -” An international research consortium today published a set of papers that promise to reshape our understanding of how the human genome functions. The findings challenge the traditional view of our genetic blueprint as a tidy collection of independent genes, pointing instead to a complex network in which genes, along with regulatory elements and other types of DNA sequences that do not code for proteins, interact in overlapping ways not yet fully understood.”

http://www.boston.com/news/glo.....ed/?page=1

“The science of life is undergoing changes so jolting that even its top researchers are feeling something akin to shell-shock. Just four years after scientists finished mapping the human genome - the full sequence of 3 billion DNA “letters” folded within every cell - they find themselves confronted by a biological jungle deeper, denser, and more difficult to penetrate than anyone imagined.”

Since everyone is performing these fantastic mathematical computations, I will simply conjur up a point backed up by third grade mathematics, and then return to my seat in the back of the class, and take the short bus home after the bell rings.

30,000 generations of e. coli, and presto, we get a new constructive function, the ability to utilize citrate. Now, 30,000 advanced primate generations, at 20 years per generation, equals 600,000 years. So, a billion years equals less than 2,000 generations.

In this drop in the chronological bucket, is it feasible that the features and abilities, particularly mental, could have evolved through undirected genetic variation and natural selection.

And for a final point, if we took our kids, surrounded them with a tiny bit of pizza and ice cream, and enormous amounts of broccoli and other assorted veggies, would they mutate fast enough to eat the veggies before the “good” stuff runs out and they starve? Or go the direction of the Great Lizards?