Questions for Sean Pitman

[Sean Pitman]

Originally posted by Jerry Smith
To establish that there was a mechanism limiting their evolutionary potential, or that something was actually blocking their evolution of a lactase enzyme, your experiment would have to run somewhat more than 20 years, or even 100 years.

Actually not.  Humans are thought to have evolved from ape like creatures in about 5 to 8 million years.  That works out to be around to about one quarter to one million generations (give or take).  Of course, bacteria have a much shorter generation time.  E. coli have a generation time of 20 minutes.  Thats about 25,000 generations per year.  So, E. coli should be able to evolve much more rapidly than creatures such as primates. 

Also, blocks to E. coli evolution, once their etiologies are determined, can be extrapolated over time to estimate how long on average it would take to overcome these blocks.  Real time experimentation is both impossible as well as impractical when the time involved runs in the thousands, millions or billions of years.  The same thing is done with radiometric decay rate calculations.  No one has directly documented the half life of Rubidium to 48.8 billion years.  This is an extrapolation based on data obtained over a much shorter course of time.

Extrapolation in the case of evolutionary blocks in bacterial evolution can also be done very reasonably.  After all, the scientific method is all about the predictive value of the hypothesis.  Without extrapolation concerning the likelihood of future events, there is no science.

I believe that the "limit" Hall mentioned was more a "limit" to what E coli could accomplish over the geological blink of an eye that was his experiment. I don't think it is appropriate to extrapolate a "limiting mechanism" from data that does not cover realistic evolutionary time frames, nor do I think Hall intended for anyone to.

Then you don't understand how science works.  The scientific method is all about extrapolation or the "predictive value of a hypothesis".  Science is all about predicting the future.  In Hall's own paper he himself made estimates about how long he thought certain mutational events should or would take.  By his own calculations he estimated that a function that required a double mutation (two neutral mutations) would take around 1 million years to evolve.  Of course he was wrong since such a double mutation was isolated in the lab in a few days.  He was mystified by this and could not explain how this mutation happened.  He even suggested that sometimes mutations are not entirely random events . . . that they are guided by some other magical process outside of the capabilities of natural selection.  He said that such a hypothesis is "heresy, I am aware", but it was the best he could come up with at the time.  Since this time higher mutation rates and genetic hot spots have been discovered that seem to explain some of Hall's dilemma, but the point remains.  Such extrapolations are not unreasonable.

After all - all of the data does prove common descent. It seems almost silly to extrapolate far outside the data to postulate a limiting mechanism exists in order to prove that something couldn't happen when we already have strong evidence that it has happened ubiquitously through history.

    - I bow to the "overwhelming" evidence.

Now that's an interesting argument.  Forget genetics and trying to figure out how evolution really works or doesn't work.  We already know that it does work because of all the other data.  So, anything that suggests that there might be a problem or a limiting factor can be overlooked.  Extrapolations based on such problems can also be overlooked.  Anything that doesn't agree with what I already know can be discounted.  You might understand "all the other data" so well that you don't need to consider any more evidence.  You have already arived at the full knowledge of truth.  However, for those of us who do not understand the "rest" of the data like you do, some of these other problems are still of some concern.

You have only lived a very short time in comparison to known history.  And yet, you extrapolate your experience to "understand" vast spans of history.  You know what happened millions and even billions of years ago based on a relatively tiny fraction of experiencial time and potential data.  And yet, you tell me that it is silly to extrapolate?  Hello!  What are you doing? 

Your argument is very much like those arguments generally used by fundamentalists of any faith group.  You have your "Bible" of truth.  Anything that goes against what this selection of data tells you must be wrong and therefore you don't have to consider it or explain it. 

 

[Jerry Smith]

Sean, if there could be a limiting agent correctly extrapolated, it would definitely be worthwhile to study it & to find out the route evolution takes around it.

My point is that his experiment was not designed to yield that kind of data & an extrapolation from his results to the idea of a limit of that sort is unjustified. After all, we are looking at one colony in the lab over what is essentially zero time, and very restricted population growth.

Forget genetics and trying to figure out how evolution really works or doesn't work.  We already know that it does work because of all the other data.  So, anything that suggests that there might be a problem or a limiting factor can be overlooked.  Extrapolations based on such problems can also be overlooked.  Anything that doesn't agree with what I already know can be discounted.  You might understand "all the other data" so well that you don't need to consider any more evidence.  You have already arived at the full knowledge of truth.  However, for those of us who do not understand the "rest" of the data like you do, some of these other problems are still of some concern.

Sure, if the problem turned out to exist, it would be interesting to study. All I said, basically, was why invent the problem in order to try to discredit a theory that has already shown its worth?

 

[RufusAtticus]

Sean,

In all your probability calculations and protein neighborhood arguments you neglect some important biological issues. An obvious one is the neglect of considerinf protein motifs. There is a lot of literature in molecular evolution about how recombination can mix and match whole protein motifs to produce novel biochemical functions. In other words, a novel gene doesn't have to be descended from just a single gene. For example, a membrane bound kinase could result from an intergenic crossover that combined the enzymatic ability of one gene, the transmembrane domain of another, and the signaling domain of a third.

Take your example of "batter." That word has multiple sources. Take the most recent one used in sports. Batter is simply the combination of the verb "bat" plus the noun suffix "-er." So if "bat" exists in the language and "-er" does too then there is no need to mutate letters/sounds/sylables to produce "batter." You just have to combine them. In this instance, "bat" and "-er" are preadaptations to the word "bat." You yourself have already admitted that three and two letter words are easy to "evolve." So you must admit that "batter" is not hard at all. Similarly the phrase "methinks it is like a weasel" is the combination of smaller segments, each with their own meaning.

You also ignore that environment determines fitness. Whatever functional neighborhood you want to calculate now for a protein is only going to apply for right now. It was different in the past and will be different in the future. Take for example the English word "goose" and the Lithuanian word "zhasis." Your argument would say that it is impossible for them to be related since they are separated by a sea of non-meaning. However, they are both descended from the proto-indoeuropean root "*g(h)ans-," which is valid in neither language. How can this be? Well the sea has shifted.

So although a sea might isolate a protein on an island, an isthmus might rise in the future and allow it to get off. More appropriately the island itself might move from point A to point B allowing the protein to make the trek also. Remember we're not talking about one island here, but say 30,000 islands each with its own trajectory. So even if isthmus formation or significant island movement is a rare event, it will still happen to a few protein neighborhoods.

 

[RufusAtticus]

Oh yeah, Sean, don't forget that you still have people to respond to in this thread.

 

[Micaiah]

Originally posted by Sean Pitman

You have only lived a very short time in comparison to known history.  And yet, you extrapolate your experience to "understand" vast spans of history.  You know what happened millions and even billions of years ago based on a relatively tiny fraction of experiencial time and potential data.  And yet, you tell me that it is silly to extrapolate?  Hello!  What are you doing? 

Your argument is very much like those arguments generally used by fundamentalists of any faith group.  You have your "Bible" of truth.  Anything that goes against what this selection of data tells you must be wrong and therefore you don't have to consider it or explain it. 

Couldn't have said it better.

 

[Jerry Smith]

And as I pointed out, you still would be incorrect - even if you found a more eloquent way of putting it.

 

[LiveFreeOrDie]

Originally posted by Sean Pitman
Certainly mutations need not produce any one particular function. The question is, given a particular genome, how many functions can it produce or is it close enough to evolve in a given amount of time? The problem is that there is a limited amount of genetic real estate available. If there were an infinite number of genetic sequences available to a given organism or pool of organisms, then the evolution of an infinite number of beneficial functions would be a simple process given the proper environment. The creatures and/or their offspring could quickly "morph" as various environmental needs arose. The problem is that there is simply not enough genetic real estate to cover the potential functions that might be useful should the need arise. Because of the limited nature of the genome (or even pools of genomes - ie: gene pools), certain functions that might be useful to a particular creature and its offspring are simply out of reach. Other functions might just so happen to be close enough to some portion of that gene pool's real estate to be enveloped by random fluctuations in that real estate. However, natural selection limits such fluctuations because it is more concerned with maintaining what is already working that it is with sending portions of its expensive real estate out on voyages across an unknown ocean. Even when such creatures do have extra real estate that can undergo random neutral mutation and "drift" along, what are the odds that they will end up at anything as far as a beneficial function? The odds are that relatively few functions are within quick reach of this relatively limited genetic real estate.

Excuse me but how does any of this pose a problem for evolution? All you are arguing is that the mechanisms of genetic mutation and inheirtance make saltational changes highly improbable. This is hardly news. Anyone with even a slight exposure to paleontology and comparative anatomy can see that nature doesn't work this way. Novel structures in living organisms are nearly always the product of exaptation.

For example, why is it that no winged vertebrates have four legs? The answer, according to what we know about evolution and genetics, is that it was simply too improbable for a population to randomly develop a completely new set of limbs and for those limbs to offer a selective advantage at the time of their presumably primitive origins.

What do we see instead? Flight evolved, but only by exapting an existing set of limbs. Careful analysis of vertebrate wing structure has identified many homologies between wings and forelimbs. Yet despite their very similar structure, wings and limbs perform very different functions.

So I'm afraid your argument is nothing but a tempest in a teapot. It's merely another attempt to raise a cloud of confusion whose sole purpose is to hide the blatant weakness of your position:

You are completely incapable of providing a testable positive hypothesis to explain the origin of species that is supported by empirical evidence.

 

[lucaspa]

Originally posted by Sean Pitman
The problem this is that every genetic function is irreducibly complex to one degree or another. Behe's particular problem is that he does not detail the degree to which the evolution of an irreducibly complex system becomes "impossible." Clearly, irreducibly complex functions can evolve.

I'm really glad you see this.  Not only that, but IC systems can easily arise by chance.  They simply don't have to be "well-matched".  This Behe also acknowledged in another post.

However, the point I wanted you to focus on was that Miller used Hall's experiment for a very limited claim.  Within those limits, Hall's experiments demonstrate exactly what Miller claims they do.

The lactase function, for example, is "irreducibly complex". This can be proven without a doubt by simply reducing the molecular structure of any lactase enzyme until it no longer works. Since not every protein can hydrolyze lactose, the ones that can are clearly irreducibly complex in their specialized function.

I want to be sure that you are aware that this isn't a quantum phenomenon where you have specific lactase activity or you have none.  You can have a gradual reduction in activity depending on the amino acid sequence.  Even to the point where you have lactase activity for both 6 and 5 carbon sugars and thus don't have a specific lactase. 

 If some series of amino acids that does not have lactase activity can be changed via mutation and natural selection into one that does have lactase activity, evolution of an irreducibly complex function has occurred. Behe makes the mistake of trying to define irreducible complexity as something more than it really is.

Yes, he does.  I'm glad you see this. 

Behe has a great idea, but fails to recognize the fact that all functions are irreducibly complex, not just the more complex functions.

I disagree.  His "great idea" is that he can recognize the action of an intelligent entity only by looking at the object.  In this case a system.  What Behe forgot was that to infer intelligent action you have to look at both the object and the environment. For instance, you are a cop investigating what might be a breakin at a house.  You have a broken window.  You can't look only at the window to determine that someone broke it.  Yes, if you have human blood on the glass shards or fibers of nylon, you conclude a human broke the window with his fist, either bare or wearing a glove.  However, if you see the wind blowing and a large branch swinging through the hole, you don't conclude intelligent action.  What Behe ignores is that chance and then natural selection in combination can also make his "well-matched" IC systems.  Therefore, he has a process in the environment that will do the job and can't conclude action by an intelligent entity.


In Hall's paper he noted, "We thus tentatively concluded that in E. coli K12 there is only one locus, the ebg locus, which can evolve a lactose-hydrolyzing enzyme under our selective conditions. That tentative hypothesis has been borne out over several years. A strain deleted for both the ebg gene and the lacZ gene has repeatedly been subjected to the standard selective regimen on Lac-TET plates, and has additionally been mutagenized with a variety of chemical mutagens before being subjected to selection. All of those experiments failed to yield any lactose-utilizing mutants. Thus, E. coli has a clearly limited evolutionary potential for this particular new function."

Hall refers to a 1973 paper by another investigator, indicating that perhaps there were 8 or 9 years that they tried.

If Hall kept a strain deleted for LacZ and ebg genes around for several years by the time this paper was published, I am betting he still keeps a colony on hand. You don't just throw away colonies of genetically altered bacteria.

Sometimes you do. If you retire or move to other work.  You don't necessarily keep the experiment running and you often freeze the bacteria down to avoid the time and expense of maintaining a live colony. Right?

Even so, I will alter the paper to reflect the fact that I know of no specific experiment has been carried out over the course of 20+ years with these bacteria.

Good. Thank you. 

Also, note that these bacteria where subjected to mutagenic chemicals that dramatically increase mutation rates . . . and they still failed to evolve the lactase function.
[/size][/font]

Obviously there were many other mutations, possibly with many other functions gained or lost. The fact remains however that the only gene that was capable of evolving the lactase function in particular was the ebg gene.

Not "capable", but did mutate in this time frame under these conditions.

No other mutation or series of mutations to any other gene or sections of genetic real estate gave rise to the lactase function despite heavy selection pressures over a prolonged course of time

Selection pressure won't cause mutations.  It only guarantees finding one if it is present.  However, the severe selection pressure can be a disadvantage in finding a new mutation.  If the pressure is so heavy that only a high specific activity will provide enough energy to keep the bacteria alive, then you will actually overlook mutations.  Mutations with lower specific activity would prove advantageous in a less severe selective environment.  So it appears that Hall may have inadvertently biased his experiment against detecting other mutations.


You evidently do not understand the experiment. The E. coli strains that had both the ebg and the lacZ genes deleted were grown on lactose selective media (Lac-TET plates). These strains did not die out because lactose selective media is not limited to lactose as the only source of energy. These colonies can be and were in fact grown on such selective media over an extended course of time, and yet they failed to evolve the lactase function despite heavy selection and mutagenic pressures.

Now, this might be a lower selective pressure, but it sounds like it is not.  Confusion. We need to check this out specifically.

Variations in the population would obviously add up over time via neutral mutations and genetic drift. And yet, none of these variations were able

The precise sentence is "none of these variations did produce".  The inference is that they were incapable, but all the data says is that the function failed to appear. 

For further validation of this concept, go to any hospital laboratory and ask to see their records on bacterial enzymatic capabilities over the past 40 or 50 years. You will note that many types of bacteria have been lactase negative over this entire time (ie: Salmonella, Proteus, and Pseudomonas etc) despite the benefits that would come if they were to ever evolve a lactase enzyme.

Good point.  However, in this case we are not looking at the total environment and the cost involved.  Without selection pressure, is a lactase clone going to have an advantage? You are assuming an advantage because of an alternative food source. However, if the food source is not abundant and the cost of producing the lactase is more in terms of ATP than the bacteria is going to get back from the lactase activity, then it is an overall disadvantage, isn't it? 

Natural selection can only work upon sequence differences if they are functionally different from each other. If they are beneficially different, they are protected or maintained by natural selection. More beneficial mutations can then be added onto this first beneficial change. This was in fact demonstrated by Hall in his experiments with E. coli. Certain of the functions that evolved via selective guidance required a series of 2 or 3 point mutations before they were realized. Of course, each point mutation was selectively advantageous in a functional way over that which came before. Thus, the evolution of such a function that took 2 or 3 point mutations was "guided" along the evolutionary path by natural selection . . . each step of the way. Without natural selection, there is no "guidance" down any path. Without the selection of particular mutations, all mutations are allowed to drift in a completely random way. Of course, a purely random process leads to homogeny, not toward increasing complexity (ie: primordial soup turning into humans).

Notice that you are supporting my contention that if the intermediate is useful, then your odds calculations don't apply.  Without selection advantage, mutations don't "drift in a random way". Instead, what happens is that Mendelian genetics and Hardy-Weinberg come into play. The frequency of that mutation in the population (number of individuals that have it) remain constant. So, you have a steady "background" of individuals that have the neutral mutation.  This frequency can increase (or decrease) by sereral other processes:  1) being linked to another gene that is selected for or against, 2) if the population is very low (<10 individuals) then random chance in the survival or reproduction by the individuals can change the frequency (genetic drift).  See Futuyma's Evolutionary Biology for the mathematical equations.

Now, of course, when the environment changes, this previously "neutral" mutation could be either beneficial or harmful because it does have a function in the new environment.

A single point mutation either does or does not result in a beneficial function. If it does result in a beneficial function, it does so instantly. There is no path that must be traversed that requires time or the "guidance" of natural selection. The function is achieved by purely random chance. Natural selection is not needed to achieve this new function. It might help to maintain the function once it is achieved, but the arrival of the function itself was purely random and sudden. A series of beneficial mutations is a guided and gradual process, but the individual mutations are not.
Consider that the protein produced by the lacZ gene is a tetramer totaling about 4,000 amino acids. Considering one unit of the tetramer (~1,000 AA), how many possible proteins of this length are there? There are 1 x 10e1301 different sequences of this length.

Now, an E. coli bacterium has around 4.1 million base pairs in its genetic real estate. ...Obviously then, the fraction of potential lactase proteins compared to the total number of possible proteins is most likely . . . miniscule

The conclusion is that the the proteins available to this bacterium  is limited, not that the total number of proteins with lactase activity is miniscule.  Humans also have lactase.  How close is it to the bacterial?  Other mammals? Do they all have the same sequences and structure? 

Think of a Library of all possible genetic sequences.  The Library of Mendel.  Down one hall are all possible sequences that will make a functional lactase, however miniscule. Now, connecting via infinite dimensions are hallways to other protein sequences.  What it appears is that these particular E. coli are in a cul-de-sac in the Library where there are no cross-corridors of sequences that E.coli has to the lactase hallway.

 

[lucaspa]

Originally posted by Sean Pitman You do understand that I set up a hypothetical situation here to think about the average time involved to cross certain neutral gaps in genetic function . . . right? The hypothetical situation that assumes a starting point of 5 neutral mutations from the lactase function would require and average time of greater than 40,000 years to cross. Of course, this is based on a large steady state population (one trillion), an extremely high mutation rate (10e-4 mutations per base pair per generation), and a fairly rapid generation time of 20 minutes.

I understand what you have done, but you ignored that bacteria have been around for 3.8 billion years.  This allows a lot of accumulation of neutral mutations within the various clones.

You suggest that there might only be one mutation needed for lactase function in a particular colony since such a sequence might have drifted along randomly over the course of the past "40,000 years" or so. You are most certainly correct. However, if such a sequence did happen to drift along to this place where only one more mutation was needed to achieve lactase function, this mutation would be realized in short order (only one generation for an average bacterial colony of 10 billion organisms).

And for a couple of Hall's clones, it was nearly this fast.

But, how often does such a sequence drift so close to one of the relatively few lactase sequences in a particular bacterial colony that could benefit from such a sequence? The fact that lacZ and ebg negative E. coli never evolved the lactase enzyme should tell us that something is blocking lactase evolution in these creatures.

And I went over that.  Energy considerations.  There is a cost to making a specific lactase enzyme, and if the level of galactose isn't sufficient to recoup that energy cost for the bacteria, any clones with lacZ is going to be at a selective disadvantage.

Why don’t they ever evolve this enzyme when it would be to their benefit if they did? 

It might even be detrimental to maintain such an enzyme since protein production uses energy resources If the protein is not beneficial, natural selection will select to get rid of it. So, at best, such mutations would come and go randomly.

SEE!!?? Sean, you are so stuck in debate mode that you are overlooking answers to your own questions.  You have the answer to the questions above right here.  This is exactly what I have been saying. As LFOD told you, instead of looking solely at the genetics, look at the fitness landscape.  Since you know this, why didn't you think of it when you asked those questions?  I submit that you were asking them only for rhetorical effect and hadn't tried to answer them.  I hope you do better when interviewing patients.

So, let me digress here: what is your emotional involvement with the subject of evolution that you are abandoning your training  and overlooking answers you already have? Perhaps if we can deal with the emotional issues, then your critical faculties will return.

 Certainly neutral mutations could accumulate that just so happen to provide lactase function before it is actually needed, as you suggest. Of course, if they did this, natural selection would not select to maintain such enzymes. . 

Unless the protein had some other function.  We get so hung up on enzymes being specific that we forget that many proteins have secondary functions. For instance, E Cabiscol and RL Levine, The phosphatase activity of carbonic anhydrase III is reversibly regulated by glutathiolation.  Proc. Natl Acad. Sci. USA93: 4170-4174, 1996. Enzyme with two active sites: an anhydrase and a phosphatase.

In any case, given a known starting point, a calculation of the average time required to achieve a particular function can be estimated.

Only if certain unrealistic assumptions are made. It is those assumptions I am disputing. 

If the particular function is more than one or two mutations away from anything currently in the genetic real estate,

This assumes that you only have point substitute mutations changing one amino acid.  Rufus quite rightly pointed out that many proteins are built upon motifs and different introns are spliced during transcription.  So then you get a whole new partial function grafted onto an existing protein. Also, point insertion or deletion mutations change more than one amino acid at a time.


I figured that eventually you would suggest that evolution is not required to travel in any one particular direction. Evolution can go in any number of directions to achieve diversity of function. Any number of different functions might be beneficial for a given organism in a given environment.

I wouldn't put it this way.  I would say rather that for any design problem, there are a number of different solutions.  Selection gets to pick among them. So having a lactase enzyme might not be necessary if there is some other enzyme that will, say, cleave galactose to a 4 carbon sugar and a 2 carbon carboxylic acid.

In reality though, such evolution might be more limited than one might think at first glance. ... Both coded systems are arbitrary.

Not quite. There is considerable evidence that the current code has selective advantages. What does that do to your argument?

1.  Alberti, S The origin of the genetic code and protein synthesis. J. Mol. Evol. 45: 352-358, 1997.

Not every series of amino acids caries meaning for a particular living system.

Actually, every sequence of amino acids will have some biological function.  Not great, and not the specific function you may be looking for, but some function.  This is another fallacy in your assumptions.

In fact, as with the English language, the vast majority of potential protein "words" are not defined by a given cell. They have no function at all for that cell. Each protein word is in fact surrounded like an island by a sea of non-defined/non-functional proteins.

Again, not true.  It may have no function for that cell but it has some biological activity.

However, there is a point beyond which proteins cannot change without a complete loss of function.

That's an assumption.  They may lose that particular function but not a total loss of function.
 C Chothia and M Gerstein, How far can sequences diverge? Nature 385:579-581, 13 Feb. 1997. Shows that you can replace quite a few residues and still have an active protein (contrary to Hoyle).  11 primary references. 

Many amino acids cannot change at all and are called "invariant".

For that function of that sequence.  However, a smaller or larger sequence can have those change and still have the same function. 

To begin looking into this question, let’s consider an average E. coli bacterium.

Instead of doing this, why don't you go to the scientific literature and see what has been done? The calculations are GIGO because you are putting garbage in.  So you get garbage out.

With all 13,000 proteins having such variation potential, the total number of different protein sequences that our bacterium could recognize would be on the order of 1 x 10e70. Certainly, this a huge number of proteins that could be recognized, but it is still a tiny fraction compared to those amino acid sequences that cannot be recognized. Subtracting the sequences that can be recognized out of the potential pool of protein sequences leaves the pool barely touched. In the pool there are still 1 x 10e130 protein sequences that have no unique function given them by the E. coli dictionary of protein words.

Which simply shows that organisms can back themselves into evolutionary corners where they cannot respond to changes in the environment.  Nothing new here.

You are treating apples and oranges.  One is the functions of proteins and the other is the variability of a particular E. coli genome. 

Of course, you can add to a genome. For E.coli this can come from incorporation of plasmids and gene duplication. 

Evolution simply cannot take off in any direction that it wants to if it plans on each mutation being functionally different from the one before.

Yes, it can. Because the new function doesn't have to be one that the E. coli has.  You emphasize "recognizes" but what that really means is that it has the same function of the proteins already in the E. coli.  So what you have just done is eliminate novelties by definition, not by data.  While the new function may not be one that the E. coli has, it could still be very useful.  An extreme example:
1.  Ohno, S, Proc. Natl Acad. Sci. USA 81:2421-2425, 1984.  Frame shift mutation yielded random formation of new protein, was active enzyme nylon linear oligomer hydrolase (degrades nylon)

Now, nylon hydrolase is not going to be at all what an E. coli might normally "recognize", but it has a function.  In this case, a very useful one.

Odds are that very soon the functional chain will dry up no matter which direction it starts off in.

That particular functional chain will dry up, but you forget that in doing this you will walk down a cross-corridor in the Library of Mendel and find yourself in a whole new hallway.  And then you have a new function to select for. 

The problem gets unimaginably worse when more complex functions (such as cellular motility or light recognition) are considered

It actually gets better, because then you can exapt whole complex functions for other uses. And then only change the Hox genes and get large changes in the organism.

The mutation rate for bacterial genomes is on the order of 1 x 10e-9 mutations per base pair per generation.

Citation? 

http://genetics.hannam.ac.kr/lecture/Mgen02/Mutation%20Rates.htm

You should read your own web site. In particular, pay attention to the difference between 'mutation rate' and "mutation frequency"  You are equating the two. For a mutation rate of 1/7 you end up with a mutation frequency of 1/4. 


"http://www.erin.utoronto.ca/~w3bio/bio370/shelley_dna_plasticity_partII/tsld014.htm"
Remembering that evolution is about populations, let's see. If we have 4.1 million base pairs per bacteria and the mutation rate is 1 X 10^-9 base pairs per generation, we only need 1,000 bacteria in a generation to have 4.1 billion base pairs and thus 4 mutations in the generation.  Right?  So where do you get that 238 generations calculation?

Even wanting only 1 out of 13,000 genes affected, that means we need 13 million bacteria to have that gene mutated in one of the bacteria in each generation, right? 

 

[Sean Pitman]

Page 1 of 2

Originally posted by RufusAtticus
Sean,

In all your probability calculations and protein neighborhood arguments you neglect some important biological issues. An obvious one is the neglect of considerinf protein motifs. There is a lot of literature in molecular evolution about how recombination can mix and match whole protein motifs to produce novel biochemical functions. In other words, a novel gene doesn't have to be descended from just a single gene. For example, a membrane bound kinase could result from an intergenic crossover that combined the enzymatic ability of one gene, the transmembrane domain of another, and the signaling domain of a third.

This is how it works in theory of course, but where has this ever been demonstrated in real time? The odds of such a gene getting mutated into just the right position so that it's kinase product combines with the signaling and transmembrane products of other genes, seems quite remote. Hybrid genes and gene products have been observed to form in real time, but all the ones that I know of cause diseases, such as leukemia/myeloproliferative disorders etc. In other words, they are not selectively advantageous in such cases.

Take your example of "batter." That word has multiple sources. Take the most recent one used in sports. Batter is simply the combination of the verb "bat" plus the noun suffix "-er." So if "bat" exists in the language and "-er" does too then there is no need to mutate letters/sounds/sylables to produce "batter." You just have to combine them. In this instance, "bat" and "-er" are preadaptations to the word "bat." You yourself have already admitted that three and two letter words are easy to "evolve." So you must admit that "batter" is not hard at all. Similarly the phrase "methinks it is like a weasel" is the combination of smaller segments, each with their own meaning.

The multiple domain argument is often used to explain the evolution of one portion of a protein through neutral or non-functional stages while natural selection maintains that protein because of the beneficial functions of another functional domains located on the same protein.

On the surface, this argument looks quite good. However, it falls apart when one realized that multiple domains are nothing more than the equivalence of two proteins stuck together. The two motifs or domains need not even work together on the same function. They can be parts of completely different systems of function. Of course, as in your word example above, they can also work synergistically.

In either case, there are problems. In the first case, two different protein functions occupying a very close proximity (ie: the same molecule) does not eliminate the problems of either one evolving a unique function. Just because one part of a protein is working does not necessarily speed up the evolution of a new function on the other domain. The working domain will be maintained by natural selection. However, neutral mutations to a second domain will continue without the guidance of natural selection. So, without guidance, they will continue to be purely based on random chance until some new function is realized. During this process, nothing about the second domain will be maintained. It will continue to mutate and drift randomly around the sea of neutral/non-function. The function of the first domain certainly doesn’t compensate for or in any way guide the random changes of the second domain.

You ask me to consider that the letter sequences "bat" and "er" are both meaningful letter sequences in the English language. You suggest that since it is easy to evolve meaningful two and three letter words, that it is also easy to put these words together to make meaningful five and six letter words. Superficially, this might make sense, until one realizes that arbitrary nature of language. It might seem reasonable that if each one of two words is defined, that making a larger word with the combination of the two smaller words, would also mean something. For example, the words "hot" and "dog" both mean something is the English language. So obviously I can put them together to make the word larger word, "hotdog" . . . and what-do-ya-know! It also means something in the English language!

Unfortunately however, this doesn’t always work. Sequences of letters that have individual meanings in English do not always have any meaning at all when they are combined. For example, the sequences "cat" and "rat" and "er" are all defined, but what does "catrater" mean? Catrater is completely meaningless in the English language. But why? After all, catrater is the combination of three other defined words is it not? Why yes it is, but the three words simply do not work together for no other reason than the fact that their union is not defined in the English language code.

Also, who is to say that when two or more sequences come together that they will come together in the proper sequence to give their combination recognized function? The sequences "hot" and "dog" could have just as easily come together to form the sequence, "hodogt". Now, what function does hodogt have?

You see, there are a lot more neutral and non-functional sequences for certain sequence lengths than there are meaningful sequences. The evolution of the meaning or function of the sequence "batter" is therefore not as simple as adding a couple of three letter words together. A six letter sequence adrift on the sea of six letter words might have to float around for a while before it happens upon any six-letter word that actually has some recognized function, much less beneficial function.

Oh, and by the way, you evidently did not understand my explanation as to why it was so easy to evolve "new" two and three letter words in the English language. You do know why don’t you? One reason is that the definition of the words was not evolved, just the pre-defined sequence. Without previous definitions being assigned to these sequences, their evolution would have been meaningless. In other words, there was already a pre-established system of function waiting for the right sequence to come along.

The other reason why such changes between words are statistically likely is that the English language defines a large majority of the total number of two and three letter words. There is no particular reason why it must be this way besides the fact that a bunch of people agreed to give arbitrary meaning to a bunch of three letter sequences that would otherwise mean absolutely nothing. The English language could easily have been set up so that only 2 three-letter words had meaning. The two defined three letter words could have been "cat" and "dog". Now, changing cat to hat to bat to bid to dig to dog would have involved four intermediate words that had no meaning at all. So, now, instead of having an intact bridge of function between cat and dog, there is a gap of non-function that separates them.

You see, for symbolic languages or codes, the ideas come first, and then symbols are arbitrarily attached to these ideas. One does not automatically evolve the meaning or function of the word "dog" just because the letters happened to come along.  Getting a sequence to having new meaning or function is a lot harder than just getting the sequence. 

End Page 1 of 2

 

[Sean Pitman]

Page 2 of 2

You also ignore that environment determines fitness. Whatever functional neighborhood you want to calculate now for a protein is only going to apply for right now. It was different in the past and will be different in the future. Take for example the English word "goose" and the Lithuanian word "zhasis." Your argument would say that it is impossible for them to be related since they are separated by a sea of non-meaning. However, they are both descended from the proto-indoeuropean root "*g(h)ans-," which is valid in neither language. How can this be? Well the sea has shifted.

I do not ignore environment or how it determine fitness. The fact of the matter is that if a mutation does not result in some benefit right now, then the current environment (nature) will not select to maintain it . . . period. Nature does not think ahead and ponder, "I wonder if this sequence might help me latter on?" No. If the sequence is not beneficial right now, nature will not maintain it regardless of its future potential. Because of this, even if a particular sequence might be exactly what would be needed in the next few generations, nature might not maintain it until then.

So, the problem comes back around to, "What are the odds that the proper sequence will be there when it is needed?" It might have come around yesterday, but if it is not here today when I need it for my current environment, what good was it if I had it yesterday when I didn’t need it?

Also, human languages and codes evolve easily . . . no doubt about it. That is because humans have the powers of high intelligence and reason. Environments do not change human codes. Humans change human codes and give the same or other arbitrary meanings to these changes. No "fitness" selection is required for the "evolution" of human language. In other words, the letters don’t just randomly mutate and the humans who see this happen think, "Aha! A new word for our language has just arrived! How fortunate to have a brand new functional word to expand our vocabulary!"

Cellular codes and the mechanical workings of DNA and proteins do not think intelligently. They cannot plan ahead, remember the past, or anticipate future needs thoughtfully. The evolution of human language is not based on random mutation and natural selection. For human language evolution, the ideas always come first and then these ideas are given arbitrary symbolic representation. There is absolutely no need for the idea of a "goose" to be represented by the otherwise meaningless sequence of letters, "G.O.O.S.E." or "Z.H.A.S.I.S." These are completely arbitrary representations of an idea. If the symbolic representation for an idea changed over time in the language systems of two groups of humans . . . well that is easy. It happens all the time. Humans can change anything they want without using symbolic mutation and natural selection.

However, cellular codes don’t work like this because they can’t "think". This does not mean that cellular information is not coded. It is. It works very much like computer code works. Computers can do some pretty amazing functions, but they simply do not evolve without the input of higher intelligence. Why not? Because computer function (phenotype) is arbitrarily attached to an arbitrary code (genotype). A change in the bits and bytes of its code (genotype) does not necessarily result in an equivalent change in its function (phenotype) regardless of the needs or pressures of its environment (me). A particular function (phenotype) might improve performance in a given environment, but if that function is attached to a sequence of zeros and ones that is more than one or two random mutations away from the sequences that are already there in the computer’s coded language system, then odds are it will take a longer time to realize that environmentally beneficial phenotypic function.

So although a sea might isolate a protein on an island, an isthmus might rise in the future and allow it to get off. More appropriately the island itself might move from point A to point B allowing the protein to make the trek also. Remember we're not talking about one island here, but say 30,000 islands each with its own trajectory. So even if isthmus formation or significant island movement is a rare event, it will still happen to a few protein neighborhoods. [/B]

Consider again how large the ocean of neutral/non-function is around these "30,000 islands". If these islands are made up of sequences of only 4 amino acids, then the 30,000 functional islands would be pretty close to each other . . . separated by only 5 neutral/nonfunctional sequences on average. However, if these 30,000 functional proteins averaged 100AA in length, each functional sequence would be separated, on average, from the nearest functional protein by a huge ocean of 4 x 10e125 neutral/non-functional proteins.

Of course, you would say that the odds would go up if the environment changed and shifted the islands around as well as adding new islands and isthmuses. Yes, this is reasonable. But how many new islands and isthmuses would an average environmental shift uncover? Remember also that an environmental shift might just as well sink a few islands and isthmuses that used to be there before. But, lets say that a shift in environment brings up more islands than it sinks by 1,000,000 to 1. This shift would give us about 30 billion islands, all in different places. On average, how much closer would this bring our islands together? Our islands would still be separated by an ocean averaging of 4 x 10e119 neutral/non-functional mutations between each one. In order for your shifting environment idea to help you much, your new hypothetical environment would have to raise trillions upon trillions of trillions more potential islands than it sank.

But maybe a growing population could enter new environments without leaving the old environments. This would raise new islands without sinking any of the ones that it had before. Let’s say that each new environment adds 30,000 new islands on average to the total number of potential islands. After infiltrating a billion new environments, the total number of potential islands would be 30 trillion. After all this spread over many different environments, the functionally beneficial islands would still be separated by an ocean of 4 x 10e116 neutral/non-functional sequences.

Explain to me again how random mutation and natural selection can cross such gaps in a reasonable amount of time? Changing environments really don’t help the odds out too much as far as I can tell. And, your appeal to the idea of human language evolution is not a true parallel in the way that you use the concept of human language evolution as an equivalent to genetic evolution. The mechanisms of human language evolution do not rely on random mutation to create new function. Human language evolution is the result of higher intelligence. It is basically evolution by design. The function itself is always there first. Old ideas are sometimes attached to new sequences, and new ideas are sometimes attached to old sequences . . . as well as new seqeunces.   

Functional genetic evolution does not have the luxury of being able to think of ideas and then attach them to sequences at will. For genetic evolution to succeed, purely random mutations must be guided by a selection mechanism that only recognized changes in function. Unlike human language, the sequence definition for various functions is set by an external source (nature). Nature is like a schoolteacher who knows the answers but will not tell the students anything until they guess it right. The students in this teacher’s class have very poor memories. They cannot remember the bad guesses of the past and repeat them over and over again. Also, these students cannot reason from cause to effect. The best that they can do is to formulate random guesses and do exactly what they are told. They cannot think for themselves and the teacher cannot teach them beyond what they currently have and what they guess by random chance. The problem is that answers to higher and higher levels of question complexity are surrounded by more and more wrong answers than right answers. There may be several correct answers, but for every potentially correct answer, there are millions, billions, and trillions of trillions of incorrect answers. For each guess, all the teacher can do is say "yes" or "no".  No other hints like, "You're getting warmer" are given out.  Since each guess takes time, the average time needed to guess the correct answer grows dramatically with the increasing complexity of the correct answers. 

Some might argue by saying that some functions can be increased or decreased in a sequential manner, one step at a time with each step being selectively advantageous.  Well of course this is true.  In my view, each and every change in function that takes place is a new and unique function.  I would even call such changes "evolution in action." 

I have no problem with the idea of evolution itself.  I just have problems with trying to understand how neutral gaps in function that are wider than one or two point mutations can be crossed by a process that is dependant upon the guidance of something that cannot see beyond these gaps.  We know that evolution cannot work by random chance alone.  That would be statistically impossible.  The guiding eye of Natural Selection is supposed to save evolution, but this eye is blinded by the gaps.  So, how are the gaps crossed?  Random chance is already ruled out, and natural selection is ruled out since it is blinded by the gaps.  So, what other process is there that crosses such gaps?  The only process that I know of that is capable of language or code evolution across such gaps is one that involves a Higher Intelligence.

Sean

 

[RufusAtticus]

Originally posted by Sean Pitman:
This is how it works in theory of course, but where has this ever been demonstrated in real time? The odds of such a gene getting mutated into just the right position so that it's kinase product combines with the signaling and transmembrane products of other genes, seems quite remote. Hybrid genes and gene products have been observed to form in real time, but all the ones that I know of cause diseases, such as leukemia/myeloproliferative disorders etc. In other words, they are not selectively advantageous in such cases.

Sean, I suggest that you read Chapter 6 in Graur and Li’s textbook, Fundamentals of Molecular Evolution, 2nd edition. You will find all the information you could want there, including references to scientific studies.

You ask me to consider that the letter sequences "bat" and "er" are both meaningful letter sequences in the English language. You suggest that since it is easy to evolve meaningful two and three letter words, that it is also easy to put these words together to make meaningful five and six letter words. Superficially, this might make sense, until one realizes that arbitrary nature of language. It might seem reasonable that if each one of two words is defined, that making a larger word with the combination of the two smaller words, would also mean something. For example, the words "hot" and "dog" both mean something is the English language. So obviously I can put them together to make the word larger word, "hotdog" . . . and what-do-ya-know! It also means something in the English language!

I have never argued that it would work every time. The point is that your probability calculations are extremely flawed if you don’t take domain shuffling into account. For example, you had previously ruled that it would be nearly impossibly to form the word “batter,” but once you take into consideration recombination the impossibility disappears.

Also, who is to say that when two or more sequences come together that they will come together in the proper sequence to give their combination recognized function? The sequences "hot" and "dog" could have just as easily come together to form the sequence, "hodogt". Now, what function does hodogt have?

The fitness criteria say. The formation of “hodogt” or “catrater” might fail, but “batter” is successful. Therefore, you can’t claim that evolution is impossible or improbable.

You see, there are a lot more neutral and non-functional sequences for certain sequence lengths than there are meaningful sequences. The evolution of the meaning or function of the sequence "batter" is therefore not as simple as adding a couple of three letter words together. A six letter sequence adrift on the sea of six letter words might have to float around for a while before it happens upon any six-letter word that actually has some recognized function, much less beneficial function.

Your criteria of this analogy require that when a protein is formed it is recognized by the vocabulary. “Batter” is a valid English word. (And in fact, even if it was not part of the vocabulary before it would be instantly recognized as such by the rules of English grammar.) Thus when it was formed it was not formed in the neutral sea of your analogy. So thus you can’t dismiss it by saying otherwise.

Oh, and by the way, you evidently did not understand my explanation as to why it was so easy to evolve "new" two and three letter words in the English language. You do know why don’t you? One reason is that the definition of the words was not evolved, just the pre-defined sequence. Without previous definitions being assigned to these sequences, their evolution would have been meaningless. In other words, there was already a pre-established system of function waiting for the right sequence to come along.

And that is exactly why “batter” is instantly functional.

So, the problem comes back around to, "What are the odds that the proper sequence will be there when it is needed?" It might have come around yesterday, but if it is not here today when I need it for my current environment, what good was it if I had it yesterday when I didn’t need it?

Sorry, but that is the teleological straw-man of evolution. The appropriate question is “what are the odds that some sequences will be around when the environmental conditions will select for them?” Evolution isn’t about the “proper” sequence getting selected for. It is about any sequence getting selected for. There are many examples in nature of selection producing adaptation, the nylon digesting enzymes are one that come to mind right now. You can calculate all your hand waving probabilities that you want. But the fact remains that we do observe evolution both morphologically and molecularly. If you still want to calculation probabilities, how about you do it rigorously? Give us the assumptions you make and use only variables. When you finally get an algebraic expression for your probability, plug in values for the variables with appropriate references to scientific literature to support the values.

Also, human languages and codes evolve easily . . . no doubt about it. That is because humans have the powers of high intelligence and reason.

There is no reason behind language change, only error.

Cellular codes and the mechanical workings of DNA and proteins do not think intelligently. They cannot plan ahead, remember the past, or anticipate future needs thoughtfully. The evolution of human language is not based on random mutation and natural selection.

You need to study historical linguistics. Language change is random. It is an unconscious process whereby mistakes made in acquisition replace the previous paradigm. Drift and migration are important parts of it. So is selection, since all language changes are not equal.

There is absolutely no need for the idea of a "goose" to be represented by the otherwise meaningless sequence of letters, "G.O.O.S.E." or "Z.H.A.S.I.S." These are completely arbitrary representations of an idea. If the symbolic representation for an idea changed over time in the language systems of two groups of humans . . . well that is easy. It happens all the time. Humans can change anything they want without using symbolic mutation and natural selection.

But symbolic mutation and symbolic selection do control language change and diversification.

However, cellular codes don’t work like this because they can’t "think".

Language change is not a thinking process either.

This does not mean that cellular information is not coded. It is. It works very much like computer code works. Computers can do some pretty amazing functions, but they simply do not evolve without the input of higher intelligence. Why not? Because computer function (phenotype) is arbitrarily attached to an arbitrary code (genotype). A change in the bits and bytes of its code (genotype) does not necessarily result in an equivalent change in its function (phenotype) regardless of the needs or pressures of its environment (me). A particular function (phenotype) might improve performance in a given environment, but if that function is attached to a sequence of zeros and ones that is more than one or two random mutations away from the sequences that are already there in the computer’s coded language system, then odds are it will take a longer time to realize that environmentally beneficial phenotypic function.

Computer code can be evolved too. The computer science field of genetic programming is doing some phenomenal work. In fact I use genetic programming to identify rules that govern the classification of data. It is much more powerful than any thinking process.

Of course, you would say that the odds would go up if the environment changed and shifted the islands around as well as adding new islands and isthmuses. Yes, this is reasonable. But how many new islands and isthmuses would an average environmental shift uncover? Remember also that an environmental shift might just as well sink a few islands and isthmuses that used to be there before.

As long as the number is greater than one, you can’t claim that it is impossible to select between two islands.

Explain to me again how random mutation and natural selection can cross such gaps in a reasonable amount of time?

Evolution doesn’t require a time limit, so placing one on it is ridiculous.

Changing environments really don’t help the odds out too much as far as I can tell.

They just happen to show that your probability calculations are bust.

And, your appeal to the idea of human language evolution is not a true parallel in the way that you use the concept of human language evolution as an equivalent to genetic evolution. The mechanisms of human language evolution do not rely on random mutation to create new function.

Yes is does, just read some of the literature on historical linguistics.

Human language evolution is the result of higher intelligence. It is basically evolution by design.

LOL. The change from Latin to French was not by design, but by mistake.

 

[RufusAtticus]

Originally posted by Sean Pitman:
Functional genetic evolution does not have the luxury of being able to think of ideas and then attach them to sequences at will. For genetic evolution to succeed, purely random mutations must be guided by a selection mechanism that only recognized changes in function.

Again you make this mistake. Selection and mutation are not the only mechanisms that cause evolution. Genetic drift, migration, and the mating system are all important features. Mutations do not have to be guided by selection for evolution to occur. Drift can accomplish evolution quite nicely on its own.

I have no problem with the idea of evolution itself. I just have problems with trying to understand how neutral gaps in function that are wider than one or two point mutations can be crossed by a process that is dependant upon the guidance of something that cannot see beyond these gaps.

Gaps are not permanent, they are temporal. So what you view as a neutral gap today might not have existed 100,000 generations ago when the protein evolved into its current state. Furthermore, not all gaps will be crossed, but drift guarantees that some will. Here is an example. A boy is in a boat on the east shore of a lake. To deicide whether he goes forward or not he tosses a coin. If it lands on heads he paddles one stroke forward. Otherwise he paddles one stroke back. If he is on the east shore and a tail comes up he goes home, similarly for heads on the west shore. There are only two places where he can stop his journey, the east and west shore. So what is the probability that he gets out on the west shore or the east shore? It is simply the 1 – (ratio of the distance he is from the shore to the total distance across the lake). (We can work through this in more detail if you need it.) If the boy does this every day of his life, then most of the days he will get out on the east shore. Some days however, he will drift across the lake and get out on the west shore. Try playing with this simulation to get an idea what drift can do. Here is a sample output; note how some drift across the gap and others do not.

We know that evolution cannot work by random chance alone. That would be statistically impossible.

Actually it is statistically guaranteed that random chance alone will cause evolution. Unless a population is of infinite size, sampling during reproduction will cause fluctuations in the allele frequencies of a population, and hence evolution.

The guiding eye of Natural Selection is supposed to save evolution, but this eye is blinded by the gaps. So, how are the gaps crossed?

Drift. See the above example. You have dismissed random fluctuations but don’t fully understand them.

 

[Sean Pitman]

Page 1 of Reply

Originally posted by lucaspa

However, the point I wanted you to focus on was that Miller used Hall's experiment for a very limited claim. Within those limits, Hall's experiments demonstrate exactly what Miller claims they do.

Miller is somewhat correct, but misses the bigger point that Behe is trying to make. Hall's experiments do not cross the multiple traffic lanes of neutral function/non-function that Behe proposes as a block to evolution. On the contrary, Hall's experiments support Behe's prediction that such blocks do exist in real life. If one million turtle start to cross a highway and only 1 of them makes it past the first lane and into the second lane, how many turtles, on average, must start across this highway before one of them makes it to the third lane? The answer is 1 trillion. To make it to the fourth lane, a million trillion turtles die trying. All of this takes time of course.

Behe’s road kill argument continues to stand. Hall’s experiments support Behe’s position more than they support Miller’s position. Hall’s E. coli crossed only one or two lanes of traffic before they got to the other side. Behe isn’t talking about crossing one or two lanes of traffic. Behe’s concern is with crossing highways that are tens or hundreds of lanes wide. Each additional lane of traffic kills off a bunch of turtles and requires that many more tries before even one turtle makes it across.

I want to be sure that you are aware that this isn't a quantum phenomenon where you have specific lactase activity or you have none. You can have a gradual reduction in activity depending on the amino acid sequence. Even to the point where you have lactase activity for both 6 and 5 carbon sugars and thus don't have a specific lactase.

Actually, it is a quantum phenomenon in this particular case. It is true that the ebg gene did produce a protein with some minimal lactase activity before mutation, but this activity was so small that it was not selectively advantageous. The single mutation isolated by Hall caused a quantum/sudden evolution of beneficial lactase function for this sequence.

It is true that protein functions in general can often be up regulated or down regulated sequentially by altering certain amino acids. Each alteration is functional however. Natural selection can influence such changes since the various degrees of function are separated by bridges in function that are only one genetic mutation wide. Such changes in the levels of a given function are no more and no less hard to evolve than radical, sudden or "quantum" changes in function that also involve bridges of function that are only one or two mutations wide. Basically, all changes in function are quantum changes to one degree or another. It is very similar to the English language in this respect. Words that are very similar in spelling may have very different functions and words that are very different in spelling may have very similar functions. It all depends on the system of function that uses or defines a particular sequence. Because of this, apparently slight changes in sequence can result in huge changes in function. Also apparently huge changes in sequence can result in quite minimal or even no changes in function. So, every change in function is really a quantum change.

For example, in Hall's experiment, he demonstrated the evolution of several different enzymatic activities to include hydrolytic capabilities on lactose, lactulose, beta-methylgalactosidase, and lactobionic acid. The lactose as well as the lactulose functions could be evolved from the same ebg gene with just one mutation each. However, the beta-methylgalactosidase required that that the evolution of either the lactulase or the lactase function had evolved first. Then, a second mutation could be added onto the first mutation to produce the beta-methylgalactosidase function. Likewise, the evolution of the lactobionase enzyme (which had no detectable enzymatic activity in the original molecule) required that the first two mutations take place first. At this point, a third mutation in series resulted in an entirely new function.

Note that each step in this sequence of evolutionary steps was each separated from the previous step by a single point mutation. Such evolution is not a problem. The problem comes when new functions are separated by gaps on neutral function or non-function where natural selection can no longer influence the changes. As long as the turtles can make it across the highway to some function on the other side, they can cross one or two lane highways forever. The problem comes when the highway itself starts having too many lanes. Turtles in the middle of the highway will not be protected until they make it off the highway. Once they are off the highway, they will be protected by natural selection. Natural selection simply cannot see them while they are on the highway. It can only see them once they get off the highway. The highways are nature’s blind spots.

Behe has a great idea, but fails to recognize the fact that all functions are irreducibly complex, not just the more complex functions.

I disagree. His "great idea" is that he can recognize the action of an intelligent entity only by looking at the object. In this case a system. What Behe forgot was that to infer intelligent action you have to look at both the object and the environment. For instance, you are a cop investigating what might be a breakin at a house. You have a broken window. You can't look only at the window to determine that someone broke it. Yes, if you have human blood on the glass shards or fibers of nylon, you conclude a human broke the window with his fist, either bare or wearing a glove. However, if you see the wind blowing and a large branch swinging through the hole, you don't conclude intelligent action. What Behe ignores is that chance and then natural selection in combination can also make his "well-matched" IC systems. Therefore, he has a process in the environment that will do the job and can't conclude action by an intelligent entity.

This is a most fascinating argument. You say that if nature can give a reasonable explanation for an observation, then an intelligent or deliberate action of a thinking mind need not be an automatic default. I couldn't agree more. The problem is, if one can't find a natural process capable of explaining what we see before us, what other option is there that explains its existence? For example, look at a Picaso painting. It might not look like much to the uninformed, but most anyone would be able to tell that it was intelligently designed. Why? Because, unlike your broken window example, there is no known natural process capable of painting like Picaso. It is relatively easy to break something, either by design or by natural accident. It is another thing entirely to create something. If your window was broken and then you came along sometime later and found that it was no longer broken, you would never in a million years think that your tree fixed your window.

The same thing is true for genetics. Natural processes can easily be found that change the function of genetics dramatically. But, these natural processes generally tend toward destruction of the genetic "structure". The windows in the genetic building are relatively easy to break. These changes do change the overall function of the genetic building. However, such changes are rather easy to explain via purely naturalistic processes. The problem comes when we see intact windows, plumbing, air conditioning, etc. when there is no naturalistic methods that we know of that can form such structures in the first place.

Natural selection is supposed to be able to save evolution from random chaos/homogeny. However, if natural selection is blind, what good is it? Nature is blind to neutral mutations. So, if neutral gaps exist between functions, how are these gaps crossed? There is only one known method by which such windows are fixed . . . and that method is deliberate design by higher intelligence.

Hall refers to a 1973 paper by another investigator, indicating that perhaps there were 8 or 9 years that they tried.

Eight or nine years for E. coli evolution are like 4 or 5 million years for human evolution.

Obviously there were many other mutations, possibly with many other functions gained or lost. The fact remains however that the only gene that was capable of evolving the lactase function in particular was the ebg gene.

Not "capable", but did mutate in this time frame under these conditions.

If we are talking absolute probabilities here, technically you are right since anything is statistically possible. However, we are talking odds here. How much time on average would such changes require? Given different average neutral distances, the time required could be anywhere from 20 minutes to trillions of trillions of years.

No other mutation or series of mutations to any other gene or sections of genetic real estate gave rise to the lactase function despite heavy selection pressures over a prolonged course of time

Selection pressure won't cause mutations. It only guarantees finding one if it is present.

Who said that selection pressure causes mutations?

However, the severe selection pressure can be a disadvantage in finding a new mutation. If the pressure is so heavy that only a high specific activity will provide enough energy to keep the bacteria alive, then you will actually overlook mutations. Mutations with lower specific activity would prove advantageous in a less severe selective environment. So it appears that Hall may have inadvertently biased his experiment against detecting other mutations.

Again, you don't seem to understand the experiment here. The selection pressures where heavily weighted in the direction of isolating lactase function. Hall was not looking for the evolution of other functions. As far as the pressure for detecting lactase function, they were not so great as to kill off the bacteria if the lactase function did not arise quickly. If it did ever arise, the selection pressures in place would quickly give a large advantage to those who possessed this function and thus lactase evolution would be quickly and obviously apparent.

You evidently do not understand the experiment. The E. coli strains that had both the ebg and the lacZ genes deleted were grown on lactose selective media (Lac-TET plates). These strains did not die out because lactose selective media is not limited to lactose as the only source of energy. These colonies can be and were in fact grown on such selective media over an extended course of time, and yet they failed to evolve the lactase function despite heavy selection and mutagenic pressures.

Now, this might be a lower selective pressure, but it sounds like it is not. Confusion. We need to check this out specifically.

Please do so. The pressures here are not "lower" per se. High selection pressure does not necessarily put a burden on the current population. The burden is placed on the population that does not evolve the helpful function AFTER this function has evolved in at least one member of the population. Once this evolution occurs, such pronounced survival advantages are given to the lucky winner that the offspring of this individual quickly replace the offspring of all the other members of that population. The speed at which the evolved advantages take over a population is what I use to estimate the degree of selection pressure for a given function.

 

[Sean Pitman]

Page 2 of Reply

Good point. However, in this case we are not looking at the total environment and the cost involved. Without selection pressure, is a lactase clone going to have an advantage? You are assuming an advantage because of an alternative food source. However, if the food source is not abundant and the cost of producing the lactase is more in terms of ATP than the bacteria is going to get back from the lactase activity, then it is an overall disadvantage, isn't it?

Yes. In this situation, the selective pressures would be against the evolution of lactase production. If it did happen to evolve by random mutation in one particular member of a colony, that member would experience a disadvantage and its offspring would be replaced by the offspring of the other members of the colony.

Notice that you are supporting my contention that if the intermediate is useful, then your odds calculations don't apply.

That is true. If all the intermediates between two different functions are useful, then there is no problem with evolution between these two different functions at all. The reason is because natural selection never goes blind as evolution travels back and forth on this particular path.

Without selection advantage, mutations don't "drift in a random way".

Actually, they do drift in a random way if there is no selective advantage. Without a selective advantage there is no way to maintain certain sequences in a population. Neutral sequences will eventually get mutated into different random sequences. They will not be maintained.

Instead, what happens is that Mendelian genetics and Hardy-Weinberg come into play.

You are somewhat right, except you might not be able to use creatures like bacteria to demonstrate Mendelian type inheritance since bacteria do not reproduce via sex (ie: genetic recombination). Also, those creatures that do reproduce with sex are not protected from the random genetic changes of those genetic sequences that are neutral as far as selection pressures are concerned.

The frequency of that mutation in the population (number of individuals that have it) remain constant. So, you have a steady "background" of individuals that have the neutral mutation. This frequency can increase (or decrease) by several other processes: 1) being linked to another gene that is selected for or against, 2) if the population is very low (<10 individuals) then random chance in the survival or reproduction by the individuals can change the frequency (genetic drift). See Futuyma's Evolutionary Biology for the mathematical equations.

You are correct if you are only looking at relatively short periods of time with low mutation rates. Certainly, once a neutral sequence is established in a population, it will be maintained for a while because of the relatively slow mutation rate as compared to genetic recombination events that could alter its presence in the population more rapidly given its proximity to other functional sequences. However, random mutations can and do affect neutral genetic sequences. Over more extended periods of time, these random mutations will completely change neutral sequences from what they were before . . . leaving no trace of their former existence. Neutral sequences are simply not protected from random change by natural selection.

Now, of course, when the environment changes, this previously "neutral" mutation could be either beneficial or harmful because it does have a function in the new environment.

This is statistically possible but very unlikely if a given sequence’s length and potential function are comparable to the size of an average gene. The odds are that if a sequence was neutral in one environment that it would also be neutral in another environment. On occasion, such as with the example of nylonase evolution, a change in environment will open up the potential for new functions that were not there in a previous environment. However, the odds for this happening drop dramatically with the increasing complexity of a given function.

The conclusion is that the proteins available to this bacterium is limited, not that the total number of proteins with lactase activity is miniscule. Humans also have lactase. How close is it to the bacterial? Other mammals? Do they all have the same sequences and structure?

If the number of proteins (with a particular function) available, to a particular individual or population is markedly limited, it can be reasonably assumed that the total number of proteins that can carry out this particular function as compared to the total number of potential proteins, is a tiny fraction as well. Just because humans and other creatures also make lactase enzymes does not mean that the proteins with lactase activity are relatively common as compared to all other potential amino acid sequences. Doing phylogenetic sequence analysis on such protein sequences is often done to estimate evolutionary divergence, but this really has nothing to do with the question at hand here. If humans and bacteria are found to have the same lactase enzyme, what does that prove? Or, if they have very different lactase enzymes, does that prove anything as far as estimating the relative percentage of lactase proteins out there as compared to the total number of possible proteins?

Think of a Library of all possible genetic sequences. The Library of Mendel. Down one hall are all possible sequences that will make a functional lactase, however miniscule. Now, connecting via infinite dimensions are hallways to other protein sequences. What it appears is that these particular E. coli are in a cul-de-sac in the Library where there are no cross-corridors of sequences that E.coli has to the lactase hallway.

Exactly. In the Library of Mendel, there are infinite sequences listed. In one tiny corner of the library is the relatively tiny collection of lactase sequences. If E. coli had access to the entire library, this would not be a problem. However, since an E. coli bacterium or colony or even the entire population of E. coli in the entire world put together has limited access to this library, odds are that what they do have access to, as far as functions are concerned, are separated by many volumes, hallways, and corridors from other sequences that might be of some benefit if they ever did get access to these other areas of the library. In order to find these sequences, the sequences that the bacteria do have must stumble blindly through the vast majority of sequences that the library does not defined as beneficial for those bacteria in any of their environments. Unfortunately, the library does not store all potentially useful sequences for a given creature and environment close to the creature or gene pool in question. The sequence functions that a given creature or pool of creatures does have access to are each separated, on average, by huge volumes of meaningless sequences from other sequences that would be of benefit. A given creature or gene pool might contain sequences from the library that are neutral as far as their situation is concerned when some other creature, somewhere else in the universe, might be benefited by that sequence if they had it. However, evolution cannot see such potential and bring needed sequences to where they would do the most good. Evolution can only work with the creature or pool of creatures at hand currently.

 

[Sean Pitman]

Page 3 of Reply

I understand what you have done, but you ignored that bacteria have been around for 3.8 billion years. This allows a lot of accumulation of neutral mutations within the various clones.

If bacteria have been around for ~4 billion years as you suggest, one would expect a whole lot of neutral sequences to have come along in this time. However, these neutral sequences do not "accumulate". They come and go. They are not saved because of the limited nature of the population size. There is a maximum steady state to any population in a given environment. If there were some way to maintain an infinite population, then this might solve some of the problems for evolution. With an infinite population, an infinite number of different sequences could be maintained somewhere in that population. But, this is not possible. Therefore, a neutral sequence that evolved a billion years ago that would have been helpful today is no longer in the gene pool because nature did not select to maintain it over the course of a billion years. For a sequence of 100AA, if a different sequence were evolved each and every year in a given lineage, it would take an average of 1 x 10e130 years for the same sequence to come around a second time.

So, the needed sequence might have evolved in any one of a number of other times and places where it was not needed, but what good does it do if it doesn’t evolve at the right time and place? It’s like shaking a bag of watch parts for a few million years. Every now and then two or three parts come together properly, but they fall apart before the rest of the parts can join them. Then, just by blind luck the rest of the parts come together sometime later. If only the first two or three parts were still together the watch would work, but they have long since fallen apart. They might come together again sometime in the future, but by that time the rest of the parts would most likely have fallen apart again.

You see, with limited access to sequences for a given environment, neutral evolution just doesn’t work. Shifting the environment doesn’t help much either since this would be like taking the watch parts from the bag and putting them in a car engine. Sure, some of the parts might help the car engine if they were arranged in the bag previously in some particular order, but what are the odds the next environment will be able to use what doesn’t work in this environment? There might be some environment somewhere that can use what is currently worthless, but what are the odds that we will move to that particular environment at just the right time (ie: before our sequence changes)? Working sequences are not the only things surrounded by islands of neutrality. Environments that might be helpful represent a tiny fraction of the total number of possible environments that would not be helpful for a given sequence.

You suggest that there might only be one mutation needed for lactase function in a particular colony since such a sequence might have drifted along randomly over the course of the past "40,000 years" or so. You are most certainly correct. However, if such a sequence did happen to drift along to this place where only one more mutation was needed to achieve lactase function, this mutation would be realized in short order (only one generation for an average bacterial colony of 10 billion organisms).

And for a couple of Hall's clones, it was nearly this fast.

It wasn’t "nearly this fast". It was that fast. Hall isolated E. coli colonies with the evolved lactase function in just one generation.

But, how often does such a sequence drift so close to one of the relatively few lactase sequences in a particular bacterial colony that could benefit from such a sequence? The fact that lacZ and ebg negative E. coli never evolved the lactase enzyme should tell us that something is blocking lactase evolution in these creatures.

And I went over that. Energy considerations. There is a cost to making a specific lactase enzyme, and if the level of galactose isn't sufficient to recoup that energy cost for the bacteria, any clones with lacZ is going to be at a selective disadvantage.

Again, you misunderstand Hall’s experiment. The experiment was set up so that if those bacteria that were lacZ and ebg negative did happen to evolve anything with lactase activity, the highly selective media (plenty of lactose) would give them a dramatic advantage over their peers. It would clearly be to the advantage of these bacteria to evolve a lactase enzyme if they could possibly do it. The fact that they never did evolve such an enzyme cannot be explained by some sort of attempt to conserve energy by not making a lactase enzyme. A lactase enzyme would have given them loads more energy. This is how Hall isolated the ebg gene in the first place. LacZ negative colonies evolved the ebg gene with just one point mutation so that it produced a lactase enzyme. Those bacteria that did produce this enzyme replaced those bacteria that did not produce this enzyme. The same thing would have happened with the colonies of lacZ and ebg negative colonies if one of them had evolved a lactase enzyme.

Really, is this something you truly have some confusion about? Hall himself said that these bacteria obviously had some sort of wall in the way of their evolutionary progress.

Why don’t they ever evolve this enzyme when it would be to their benefit if they did?

It might even be detrimental to maintain such an enzyme since protein production uses energy resources If the protein is not beneficial, natural selection will select to get rid of it. So, at best, such mutations would come and go randomly.

SEE!!?? Sean, you are so stuck in debate mode that you are overlooking answers to your own questions. You have the answer to the questions above right here. This is exactly what I have been saying. As LFOD told you, instead of looking solely at the genetics, look at the fitness landscape. Since you know this, why didn't you think of it when you asked those questions? I submit that you were asking them only for rhetorical effect and hadn't tried to answer them. I hope you do better when interviewing patients.

Iucaspa, you are taking my statements out of context and quoting only portions of paragraphs together with entirely separate statements in order to build your straw man representation of my position.

What I originally said was:

" Certain neutral mutations could accumulate that just so happen to provide lactase function before it is actually needed, as you suggest. Of course, if they did this, natural selection would not select to maintain such enzymes. It might even be detrimental to maintain such an enzyme since protein production uses energy resources.."

When given the entire context of my statement, I am obviously talking about a the production of an enzyme that would not benefit a bacterium in a particular environment. In the environment of lactose, the production of a lactase enzyme would be a benefit. So, the failure of Hall’s double mutant bacteria to produce a lactase enzyme in such an environment, cannot be explained by saying that "the production of lactase would not have been a benefit to them" . . . because it would have been a benefit to them. You do see that don’t you? Hall certainly did. That is why he referred to these bacteria as having, "limited evolutionary potential."

We get so hung up on enzymes being specific that we forget that many proteins have secondary functions. For instance, E Cabiscol and RL Levine, The phosphatase activity of carbonic anhydrase III is reversibly regulated by glutathiolation. Proc. Natl Acad. Sci. USA93: 4170-4174, 1996. Enzyme with two active sites: an anhydrase and a phosphatase.

Proteins with two or more domains are often used to explain how one part of a protein can evolve through neutral territory while being protected by natural selection due to the presence of another active site or domain that does have a selectively beneficial function. The problem with this idea is that a protein with two are more domains is basically the same as two or more separate proteins stuck together. Just because one domain has a beneficial function that is maintained by natural selection does not mean that any guidance is gained for the evolutionary turtles of the other domain as they try to cross a multi-lane highway of neutral function. The sequence of the neutral domain will not be protected any more than if this sequences was a completely separate molecule. If anything, the free evolution of a neutral domain is even more limited in where it can and can’t go by the fact that it is part of a larger molecule that has one or more other domains that do have functions. A major conformational change to a neutral domain could destroy the functions of the other domains. This therefore limits the types of changes that can occur even in the functionally neutral domain on its trek toward new function.

 

[Sean Pitman]

Page 4 of Reply

If the particular function is more than one or two mutations away from anything currently in the genetic real estate,

This assumes that you only have point substitute mutations changing one amino acid. Rufus quite rightly pointed out that many proteins are built upon motifs and different introns are spliced during transcription. So then you get a whole new partial function grafted onto an existing protein. Also, point insertion or deletion mutations change more than one amino acid at a time.

Sticking various proteins onto other proteins to get a combined function is like trying to stick different words together to get a combined function. The odds that the words will stick together to produce a neutral or even detrimental combined function are, on average, astronomically higher than the odds that they will yield a beneficial combined function. Also, the combination may come together in such a way that both functions are destroyed.

Also, insertions or deletion mutations may cause frame shifts that change the resultant code for many amino acids. However, the odds that such frame shifts will produce something beneficial get worse as the complexity of a particular function increases. For example, a frame shift mutation was responsible for the nylonase function discovered in certain bacteria in a pond behind a nylon factory. The fact of the matter is, the nylonase function is relatively simple. Other functions are a lot more complex, such as motility or vision. It would be impressive indeed if a non-motile bacterial colony suddenly evolved the motility function after a frame-shift mutation. Why would this be so much more impressive? Because, the relative rarity of the total number of combined sequences needed to produce any kind of motility is extremely small (much smaller than that needed for simple enzymatic activities such as lactase or nylonase) when compared to the total number of sequences that would wouldn’t produce anything, much less motility.

I figured that eventually you would suggest that evolution is not required to travel in any one particular direction. Evolution can go in any number of directions to achieve diversity of function. Any number of different functions might be beneficial for a given organism in a given environment.

I wouldn't put it this way. I would say rather that for any design problem, there are a number of different solutions. Selection gets to pick among them. So having a lactase enzyme might not be necessary if there is some other enzyme that will, say, cleave galactose to a 4 carbon sugar and a 2 carbon carboxylic acid.

Finally, perhaps we are getting somewhere. Certainly there are many different ways to skin a cat. The problem is that comparatively speaking, there are a lot more ways that wouldn’t work at all for the job of skinning cats. How does one just happen upon one of the many correct choices if each of these correct choices is surrounded by a mountain of choices that won’t work?

You say that there are quite a few different ways that one could use lactose for some benefit without having to resort to a lactase enzyme. This is certainly true, but given all these potential uses for lactose, these are still vastly outnumbered by those amino acid sequences that would give no benefit at all in their interaction with a lactose environment. Each beneficial sequence is surrounded by a ocean of non-beneficial sequences. Hall’s experiment supported this conclusion. If even one of the double mutant E. coli had found any way whatsoever to use the surrounding lactose in their environment to some advantage, any advantage whatsoever, then natural selection would have aided that bacterium and its offspring so that over the course of time its offspring would replace those without this mutation. This never happened. Plain and simple, this never happened. What does this mean? It means that even if there are many different potential sequences with beneficial function, it takes a long time to find them if they are more than one or two point mutations away from what is already there.

In reality though, such evolution might be more limited than one might think at first glance. ... Both coded systems are arbitrary.

Not quite. There is considerable evidence that the current code has selective advantages. What does that do to your argument?

1. Alberti, S The origin of the genetic code and protein synthesis. J. Mol. Evol. 45: 352-358, 1997.

There is no reason why any particular codon must code for any particular amino acid. This is purely arbitrary. The same is true for the Morse Code, the English language, or any other coded system of function. Codes are, by nature, arbitrary. The symbols within the codes carry no inherent meaning. Their function is much higher than any information they themselves contain. Some, such as Alberti, might suggest that the three-lettered structure of the codon is ideal for the needs of living things, but this is based on the fact that this is the way things already work. Who is to say that two or four letter words couldn’t have done the job just as well? What about the number of letters? Why four? Who is to say that four letters definitely work better than a 2 or 3 or 5 letter alphabet? Perhaps there are a few good reasons for the current structure of the genetic code, but the fact remains that any number of other ways could also have worked out just fine. In fact, not every living thing has the same genetic code. Some living creatures have codons that code for different amino acids than the same codon in another creature. The code is arbitrary.

Not every series of amino acids caries meaning for a particular living system.

Actually, every sequence of amino acids will have some biological function. Not great, and not the specific function you may be looking for, but some function. This is another fallacy in your assumptions.

Actually, you are technically correct. However, not every series of amino acids caries new meaning for a given living system. Also, only the tiniest of fractions of the total number of possible amino acid sequences caries beneficial meaning for a particular living system. Is that better?

In fact, as with the English language, the vast majority of potential protein "words" are not defined by a given cell. They have no function at all for that cell. Each protein word is in fact surrounded like an island by a sea of non-defined/non-functional proteins.

Again, not true. It may have no function for that cell but it has some biological activity.

If a protein has "some biological activity" then it must have it where it is found. If it has activity or function only in certain situations, but none of those situations are found in the same location currently, then, for the cell that this protein is in, the protein is worthless. Again, if a protein is not beneficial for the current cell that it is in, natural selection will not select to maintain it in that cell. If it is not maintained, the DNA sequence coding for it will mutate to something else randomly. Then later, if something comes along that would have made such a sequence beneficial, it is too late . . . the sequence is already gone.

However, there is a point beyond which proteins cannot change without a complete loss of function.

That's an assumption. They may lose that particular function but not a total loss of function.
C Chothia and M Gerstein, How far can sequences diverge? Nature 385:579-581, 13 Feb. 1997. Shows that you can replace quite a few residues and still have an active protein (contrary to Hoyle). 11 primary references.

Many amino acids cannot change at all and are called "invariant".

For that function of that sequence. However, a smaller or larger sequence can have those change and still have the same function.

Of course you are right. If an "invariant" sequence changes the current function will be lost, but that does not necessarily mean that all function will be lost. Some other function may come along that replaces the previous function. Of course, you forget that functions must be beneficial as well. The odds that a sequence, chosen at random, will be both recognized as functional as well as beneficial, in a given cell, are extremely remote (and getting more and more remote with increasing size/complexity).

On average, beneficial proteins are in fact surrounded by huge oceans of potential proteins that would not be beneficial if produced. How does one cross this ocean to something that would be beneficial? I keep asking this question and you keep giving me the run around. Your answers simply don’t explain how these gaps in beneficial functions are crossed.

With all 13,000 proteins having such variation potential, the total number of different protein sequences that our bacterium could recognize would be on the order of 1 x 10e70. Certainly, this a huge number of proteins that could be recognized, but it is still a tiny fraction compared to those amino acid sequences that cannot be recognized. Subtracting the sequences that can be recognized out of the potential pool of protein sequences leaves the pool barely touched. In the pool there are still 1 x 10e130 protein sequences that have no unique function given them by the E. coli dictionary of protein words.

Which simply shows that organisms can back themselves into evolutionary corners where they cannot respond to changes in the environment. Nothing new here.

Really? Nothing new here? Well then, explain to me exactly what keeps these organisms in these corners? What if the very first organism started out in an evolutionary corner and couldn’t respond to changing environments? Where would we be today?

 

[Sean Pitman]

Page 5 of Reply

You are treating apples and oranges. One is the functions of proteins and the other is the variability of a particular E. coli genome.

Of course, you can add to a genome. For E.coli this can come from incorporation of plasmids and gene duplication.

Plasmids and gene duplication are the answers to the problems here? Plasmids can and do bring new and functional information into many different types of bacteria. However, the information contained in a plasmid was not evolved by the bacteria that gained that information. Therefore, gain of information via plasmids really isn’t an example of evolution in action. So, that explains nothing.

Gene duplication though. Now that is an interesting and very popular concept. Let’s look at it a bit closer though. If a gene gets duplicated in a genome, there would be two genes coding for identical information where before there was one. Of course the idea is that one of these can then go on to mutate and evolve some other function for the cell while the other gene keeps the same function and is protected by natural selection. The problem here is that once the second gene gets mutated into a neutral or non-functional state, it has the same problems finding a new function as any other genetic sequence does. It is adrift on the ocean of neutrality/non-beneficial functions that separates potentially beneficial functions from each other. The time involved is no different. If it was bad before, it is still bad with extra genes.

You might argue of course that more genes means more genetic realestate and thus more chances to evolve one of the rare beneficial sequences. This is true, but compared to the massive amount of garbage sequences out there, finding the few beneficial sequences isn’t going to be much improved even if the genome expanded by trillions of genes.

Evolution simply cannot take off in any direction that it wants to if it plans on each mutation being functionally different from the one before.

Yes, it can. Because the new function doesn't have to be one that the E. coli has. You emphasize "recognizes" but what that really means is that it has the same function of the proteins already in the E. coli. So what you have just done is eliminate novelties by definition, not by data. While the new function may not be one that the E. coli has, it could still be very useful. An extreme example:
1. Ohno, S, Proc. Natl Acad. Sci. USA 81:2421-2425, 1984. Frame shift mutation yielded random formation of new protein, was active enzyme nylon linear oligomer hydrolase (degrades nylon)

You misunderstand me. I am not at all suggesting that only those sequences that an E. coli bacterium already has are the only ones that it will recognize. By all means no. Have you not been hearing anything that I’ve been saying? If something "evolved" that was already there, where is the change? Obviously living systems could use many things that they don’t have. LacZ negative E. coli can use and benefit from the evolution of another lactase enzyme. Likewise, lacZ and ebg negative E. coli could also benefit from a lactase enzyme if it were ever to evolve in any one of them.

Evolution can’t go in any direction from a given starting point if that starting point is completely surrounded by sequences that carry no beneficial function for a given cell. The fact of the matter is, there are far far far more non-beneficial sequences than there are beneficial sequences from the perspective of a given cell. An average gene in a given cell is surrounded by vast oceans of potential genes that are not beneficial. These oceans separate the potentially beneficial sequences from each other.

Now, nylon hydrolase is not going to be at all what an E. coli might normally "recognize", but it has a function. In this case, a very useful one.

Certainly an E. coli bacterium in a nylon rich environment would easily "recognize" a nylonase enzyme. The energy that is produce from the breakdown of nylon is of a type that E. coli are already set up to utilize. However, if the nylonase enzyme happened to evolve in an E. coli bacterium that was not exposed nylon in its environment, then the nylonase enzyme would be recognized as detrimental, and would not be protected by natural selection.

Odds are that very soon the functional chain will dry up no matter which direction it starts off in.

That particular functional chain will dry up, but you forget that in doing this you will walk down a cross-corridor in the Library of Mendel and find yourself in a whole new hallway. And then you have a new function to select for.

A bacterium walks down the cross-corridor in the Library of Mendel and trades in one of its neutral sequences for another sequence at random. What are the odds that this trade will result in a beneficial function? Very very very low. Odds are that this bacterium will walk down a whole lot of corridors and try out a whole lot of sequences before it every comes across one that will be functionally beneficial. In fact, it may take trillions upon trillions of years for this to happen, depending upon the complexity of the sequences and functions involved.

The problem gets unimaginably worse when more complex functions (such as cellular motility or light recognition) are considered

It actually gets better, because then you can exapt whole complex functions for other uses. And then only change the Hox genes and get large changes in the organism.

Oh really? Please explain? The more complex the function, the easier and therefore the faster it is for this function to evolve? Give me a break! By this logic, lacZ / ebg negative bacteria should evolve light sensing capabilities before they evolve a lactase enzyme. Now your are just being outright ridiculous. You are so reaching for straws here.

 

[Sean Pitman]

Page 6 of 6 Reply

The mutation rate for bacterial genomes is on the order of 1 x 10e-9 mutations per base pair per generation.

http://genetics.hannam.ac.kr/lecture/Mgen02/Mutation%20Rates.htm

You should read your own web site. In particular, pay attention to the difference between 'mutation rate' and "mutation frequency" You are equating the two. For a mutation rate of 1/7 you end up with a mutation frequency of 1/4.

Where do you get that I am equating the two? The mutation rate for E. coli is on the order of 1 x 10e-9 mutations per base pair per generation. That is the rate, not the frequency. Not that the mutation frequency in an expanding population is more than the mutation rate in that population. However, the mutation frequency in a stable population depends upon the nature of the mutation. In bacteria, if the mutation were a neutral mutation, then the frequency of that mutation in a stable population would be 1 over whatever the size of the population was. If the mutation were a beneficial mutation, then the mutation frequency would expand until it reached 100%. Note that this is different from a neutral mutation in an expanding population where the percentage of those with the mutation in the population stays the same as the percent that the first member occupied within that population. For example, if the neutral mutation occurred in one cell when there were just two cells, the expansion of the population would increase the total number of individuals, but it would not increase the percentage of individuals with the mutation. The mutation frequency would remain at 50%. If the mutation happened at the 4 cell stage, the frequency percent would remain at 25% . . . etc.

So, please do review this information again before you inform me that I am equating mutation rate with mutation frequency.

Remembering that evolution is about populations, let's see. If we have 4.1 million base pairs per bacteria and the mutation rate is 1 X 10^-9 base pairs per generation, we only need 1,000 bacteria in a generation to have 4.1 billion base pairs and thus 4 mutations in the generation. Right? So where do you get that 238 generations calculation?

It’s a lineage calculation. If you take a particular bacterium out of a population, how many mutations will its genome have when compared with the genome of its great great . . . grandfather some 1,000 generations back? With a mutation rate of 1 x 10e-9 per base pair per generation and a genome size of 4.1 million (not billion) base pairs, this genome would have to get reproduced about 240 times before one mutation would show up. So, you then ask, how many reproductions did it take to go from the grandfather cell to the grandson cell some 1,000 generations down the line? Obviously, it took 1,000 reproductions. 1,000 divided by 240 = about 4 mutations in the grandson bacterium.

Even wanting only 1 out of 13,000 genes affected, that means we need 13 million bacteria to have that gene mutated in one of the bacteria in each generation, right?

Not quite. With a genome of 4.1 million base pairs, the average gene size of 13,000 genes would be around 300bp. If one base pair has a 1 x 10e-9 chance of getting mutated with each generation, than a series of 300bp has a 3.3 x 10-6 chance of getting mutated with each generation. So, it would take around 3.3 million bacteria in each generation to get at least one mutation in every gene.

So, I propose that the odds of getting one, or even two necessary mutations "right" in a reasonable amount of time, even in a gene some 3,500 base pairs in length, is not unimaginable . . . since this gene is starting very close to the "goal" to begin with. As a hypothetical example, consider the following parameters:


 

<DIR>· A stable population of one trillion (1 x 10e12)<SUP>&nbsp;E. coli bacteria (equal to the volume of 10mL or an average hospital syringe)

· A bacterial generation time of 20 minutes (E. coli)

. An average bacterial genome of 4.1 million base pairs (E. coli)

· A mutation rate of 1 x 10-4 mutations per base pair per generation. (Hall proposes that bacteria undergo "hypermutation" in starvation states. The average rate is much lower at about 1 x 10-9 mutations per base pair per generation) </DIR>

Given these parameters, how long would it take to evolve the first mutation? The odds that the correct position would be mutated in a clonal E. coli population would be 1 in 4.1 million per genome. The odds that the correct base would get mutated at this location are 1/4. The odds that both would happen are 1 in 16.4 million mutations. With a stable population of one trillion and a mutation rate of about 410 mutation per genome per generation, the "correct" mutation will occur in about 25,000,000 bacteria in the first generation.

But, what if the starting sequence was two mutations away from the desired lactase function? Would that make any difference? The odds of getting either one of the two needed mutations with the first mutation are 1 in 8.2 million mutations. The odds of getting the second mutation right are 1 in 16.4 million mutations. So, the odds of getting both mutations right are 1 in 134,480,000,000,000 (~1.34 trillion) mutations. With these odds, the correct mutation will occur in about 315 bacteria in the first generation.

If the starting sequence was three mutations away from the desired function, the odds of getting all three mutations right shoots up to 1 in 735,157,333,333,333,333,333 (~735 million trillion) mutations. With these odds, the correct mutation will occur in one bacterium in 1,792,682 generations. With a generation time of 20 minutes, that's a bit over 68 years.

Four mutations would take about 279,740,233 (~280 million) years.

Sean

End page 6 of 6 Reply</SUP>

 

[RufusAtticus]

Sean,

Didn't your math teacher every teach you to show your work? You've tossed alot of numbers at us, but I see you offering no justification for them. If they come from the scientific literature, please give us a reference. If they are of your own divising please show us your methodology. If you need an example, look at my work in the mutation thread, which you apparently accept since you haven't been back there.