Punctuated Equilibria theories

[bumab]

Yeah we are. Your turn to straighten us out

[Biochemist]

Well, I reread a couple of your posts and I think it comes down to this:

Originally Posted by Buffy

...
Now ask me the difference inprobabilities given these two situations:

• A bunch of pieces of metal, screws, bolts sitting in a pile with whizzing things all come together at once to form a carburetor.
• A carburetor is sitting there and gets another bore drilled into it, and an intake assembly plops into it and gets connected to the same linkage that operates the others.

According to your brand of math, these are equally likely. They really aren't. Honest. Think about it.

The problem with your second example is that newly expressed genes cannot be addressed by your "Bayesian" approach. You (and Bumab) have often referenced (correctly) that prior genes are often carried forward inactively into future daughter species. But the problem is the inverse. There are many genes in daughter species that apparently never existed before. They are essentially genes de novo. There is no proffered "mutative" mechanism for a newly expressed gene (de novo-1000 bases) since it could not be "selected" without being expressed phenotypically. If you hypothezise that the internal cell machinery steers the new gene toward expression in the absence of phenotypical expression, what role did "mutation" play in this? I suggest that the answer is "none".

And this:

...I'm pointing to literature along the way here that shows various mechanisms that show how mutations can be propagated at least. ...

Showing that a DNA change can be propagated does not show that it was caused by mutation. We have a number of examples of demonstrable changes in DNA code, and I do NOT think they are mutations, because their behavior has such high specificity. Specifically:

1. Introns increase with sophistication of phyla. Bacteria/prokaryotes have essentially none. Single celled eukaryotes have very few. Higher phyla have increasingly more. But introns are not randomly inserted into genes. They are exquisitely accurate in their inclusion in a gene, and in their excision from the immature messenger RNA. It is a stretch to suggest that introns show up randomly in genes, and then concurrently generate their own mechanism to cut themselves back out of the mRNA. They are transcribed into mRNA and then cut back out. We have not found what those excised RNA fragments do yet (anyone?) but I am sure they have a purpose, because there are so few examples of the cell wasting energy to create things only to subsequently destroy them. Hypothesis: These are not mutations, they are a proscribed genome expanding mechanism, and they have an as-yet-undiscovered intracellular purpose.
2. Transposons also have specific insertion sites in their targets. They can perform either cut and paste operations or duplication operations at target receptors. Hypothesis: These are not mutations, they are a proscribed genome expanding mechanism.
3. The notion that a protogene can be created incrementally in a dormant state and then expressed en masse is starkly at odds with any of the arithmetic we know about the cell. In humans there are 3.5 billion bases available that are not part of genes. I suspect that almost ALL of this is functional, but we do not know that yet. Let's pretend for a moment that it is all "junk". That would be fodder for about 3.5 million typical 1000 base-pair genes "lying in wait". There is absolutley no credible mechanism to get a single live gene out of 3.5 million genes in only 10^14 generations (assuming 20 minute generations since the first prokaryote) since there are about 10^600 base pair combinations for each gene. If you want to argue that genes were selected, that only applies where they were expressed phenotypically. The genes that I am focusing on are the new ones (remember we are talking about PE) and there is no way for a new gene to arrive unless it was proscribed.
4. Cross-over of chromosomes is generally very explicit. There are sites where crossover is likely, and sites where it is not.

Overall, no one is agruing against mutation. We are discussion whether mutation has a significant role in speciation, and in particular, speciation related to PE. The speed of speciation in PE argues for a very structured process for speciation. Given the incredible numbers, it is difficult to argue that mutation has any hand in gradualism at all, since there have not been enough generations since the first prokaryote to create a single 1000 codon gene de novo without intermediate phenotypical expression to "steer" it. Even with some sort of selective steering, it is still a tough arithmetic hill to climb. And that is only for a single gene. But it is even more difficult to argue that mutation is a primary driver of speciation at those points in history where major phyla pop up on the paleontological record without prior phenotypical expression.


The biochemical evidence points toward programmed specificity of changes in the genome, not any random activity.

[Buffy]

Originally Posted by Biochemist

The problem with your second example is that newly expressed genes cannot be addressed by your "Bayesian" approach. You (and Bumab) have often referenced (correctly) that prior genes are often carried forward inactively into future daughter species. But the problem is the inverse. There are many genes in daughter species that apparently never existed before.

Now given that we've barely scratched the surface on comparitive gene analysis between species, and as we all seem to agree, the junk DNA that has been ignored in the past is proving important, making this sound like a definitive conclusion is a tremendous stretch! People have only recently been starting to look for familiar stuff in the junk areas, and they're finding all sorts of interesting things about certain genes replicating themselves all over the place in the middle of the junk. Moreover, because of the delicacy of DNA we only have comparisons between separate branches rather than comparisons to the root, and there's every reason to believe that "missing" elements in an apparent "parent" branch have evolved as well.

More importantly though, you are explicitly assuming something that Bumab and I are not:

Originally Posted by Biochemist

They are essentially genes de novo. There is no proffered "mutative" mechanism for a newly expressed gene (de novo-1000 bases) since it could not be "selected" without being expressed phenotypically.

Just because its an intron, does *not* mean it was not always so, in fact there's growing evidence now that the elements get copied all over the place, giving lots of opportunities to be expressed.

Originally Posted by Biochemist

I suspect that almost ALL of this is functional, but we do not know that yet. Let's pretend for a moment that it is all "junk".

Why assume this? Its called junk, but it sounds like we all agree that its all "functional-but-inactive". Since you base all your computations on the notion that the junk is completely random--the *only* situation where it would be valid--the computations are simply not valid.

Reading this post though, its clear that you do think that there's something about taking a functional but inactive intron and mutating a single base pair within it that is under all circumstances inconsequential. You have agreed that mutations happen, that's progress, but you have not explained why you believe they have no significant effect.

I think this goes back to the first point in this post of believing that we are arguing that its all junk, and while you later qualify that you are talking about "new genes", you're still insisting that they cannot be created from incremental changes. "New genes" however seem to be in they eye of the beholder: you are insisting that they are entirely new sequences never seen before created instantly. It does appear however that slight changes in sequences can have huge effects morphologically, and in any case, the existence of all those introns makes it possible for large numbers of them to sit around, having been turned on in the past to ensure their viability, but be turned on in a great wave due to environmental stress, a process that leads to another disagreement here.

You seem to be arguing for a logical switching on of certain genes under stress leading directly from one parent to one daughter branch. In fact however, if stress induces these changes (I think we agree on that) there is no reason to believe that only one line of new genes would be expressed, in fact its much more likely that many different newly expressed genes would show up in different offspring, in which selection would prove which are beneficial.

Originally Posted by Biochemist

Overall, no one is agruing against mutation. We are discussion whether mutation has a significant role in speciation, and in particular, speciation related to PE. The speed of speciation in PE argues for a very structured process for speciation. Given the incredible numbers, it is difficult to argue that mutation has any hand in gradualism at all, since there have not been enough generations since the first prokaryote to create a single 1000 codon gene de novo without intermediate phenotypical expression to "steer" it.

Well in order for this to be true you'd need to insist that somehow the mechanisms for copying exons are different than copying introns so that no steering could occur without expression. While RNA strongly emphasizes copying exons (the SciAm 4/05 article shows we now know they copy introns too), the copying mechanisms are roughly the same and do the same "steering" to both types of genes: stuff that is not copied could be radically different, but a lot would not pass too, and thus it is a *highly* selective process, making random combinatorial probability irrelevant. You keep insisting that the new genes have to be constructed from whole cloth, while its clear that there is a highly structured process--that I think has been evolved, not preprogrammed--that can easily be explained by serial gradual mutation that is expressed in a punctuated fashion:

Originally Posted by Biochemist

But it is even more difficult to argue that mutation is a primary driver of speciation at those points in history where major phyla pop up on the paleontological record without prior phenotypical expression. The biochemical evidence points toward programmed specificity of changes in the genome, not any random activity.

And there's the basic crux of the matter: if there's any programmatic process in evidence, it must dominate the mutations. Guess what? We don't disagree that the programmatic functions built into the process *make it work*, but you insist that mutation can have *no significant role* and that the process itself *could not have evolved*. I think it is still up to you to explain why mutations would have no effect in this process: You agree they occur, but you insist that either any incremental changes must be expressed immediately (obviously invalidated by evidence of PE) or they have no effect. We agree they are not expressed immediately, we disagree on the nature of those changes: you insist that the only way to create "new genes" is via the programming, and the programming ensures that no mutations *ever* affect the construction of the new gene. You berated me in an earlier post about how unbelievably smart the self-correction mechanism would have to be, but your theory in fact requires a much more strict and complex mechanism: only-good is copied with pre-programmed changes added in appropriately, not dont-copy-known-bad and try them out over time. This mechanism does presuppose that the programming knows what the next logical step is to code the new gene, but there's no obvious mechanism for that, and as I keep pointing out and you keep avoiding, it assumes that *all* future evolutionary changes are in there somewhere too--just saying they're "programmed" rather than pre-existing does not solve the information-theory related problem that they have to be in there somewhere. On the other hand, mutations provide input of new "concepts" into the system that do not have to be "pre-programmed", and with outside input, we get away from the information theory problem that you refered to in your orignal post of "loading". Yes it is a problem, and its the 500-pound gorilla standing behind you.

Bottom line: if you're going to argue that they have no effect on major morphological/phenotypic changes, while at the same time arguing that they do make changes to genes, you need to explain how they are forever bypassed. Your logic on disproof of mutations is questionable without this explanation, and unless you show they are bypassed, your math continues to be wrong and does not provide convincing proof.

Cheers,
Buffy

[bumab]

Well said, Buff.

I'll leave my theories out of it until we get this straightened out.

[lindagarrette]

Originally Posted by Buffy

it possible for large numbers of them to sit around, having been turned on in the past to ensure their viability, but be turned on in a great wave due to environmental stress, a process that leads to another disagreement here.

Buffy, Your explanation makes good sense to me except for the above statement. How could inactive (junk) DNA combinations be "turned on" for some members of a species and not for others? What would cause the selection? Linda

[Biochemist]

Originally Posted by Buffy

Now given that we've barely scratched the surface on comparitive gene analysis between species, and as we all seem to agree, the junk DNA that has been ignored in the past is proving important, making this sound like a definitive conclusion is a tremendous stretch!

First, I was being hardly definitive. Second, I am not sure which is the stretch: believing that genes that preexisted suddenly launched new dramatic phenotypes (without evidence, and against probability) or presuming that the parent species had a prespecified tendency for change in the daughter species (with partial evidence).

...Moreover, because of the delicacy of DNA we only have comparisons between separate branches rather than comparisons to the root, and there's every reason to believe that "missing" elements in an apparent "parent" branch have evolved as well.

I don't know what you mean here, Buff.

...More importantly though, you are explicitly assuming something that Bumab and I are not...Just because its an intron, does *not* mean it was not always so...

My argument is only tangentially related to introns here. At PE events, there are sudden expressions of phenotypical changes that are dramatic enough to drive classification of new phyla. It would be reasonable to assume that at least some of the (at least) several hundred new genes expressed in the 100 or so new phyla that presented at the beginning of the Cambrian period were "new" in that they were not previously phenotypically expressed. Even if some of their code was used previously, that does not really solve the mathmatical problem unless the vast majority (certainly higher than 99.9%) of each gene was used previously.

Further, the problem is not just phyla. Per the following link: http://id-www.ucsb.edu/fscf/library/.../cambrian.html , the same problem occurs with nearly all animal orders as well. We see sudden speciation regularly.

It would be truly a stretch to suggest that significant portions of the functional, coded elements of the genome that have had no previous opportunity for phenotypical steering via natural selection, were "retained" and reused suddenly- unless the genome had a prescribed mechanism to do it.

in fact there's growing evidence now that the elements get copied all over the place, giving lots of opportunities to be expressed.

Yes, and the targets for copying are all specific. I am suggesting these are intended copies, not mutated copies.

Why assume this?

I was trying to improve the probability for your position before I described that it was highly improbable under overly favorable circumstances.

...Since you base all your computations on the notion that the junk is completely random--the *only* situation where it would be valid--the computations are simply not valid.

Sure, but I was trying to make YOUR case more favorable. If you assume that the majority of "junk" DNA is actually functional, your probabilistic case gets substantially worse. I suspect that nearly NONE of the "junk" DNA is nonfunctional, but that is only a suspicion, based on extrapolating the increases in knowledge of the genome over the last 50 years. If NONE of our genome is "junk", then it makes the random mutation argument even more untenable.

...You have agreed that mutations happen, that's progress, but you have not explained why you believe they have no significant effect.

It is that nasty evidence problem. There is not ANY evidence that mutation drives speciation. There is quite a bit of evideence that mutaion causes disease, but not speciation.

...It does appear however that slight changes in sequences can have huge effects morphologically, and in any case, the existence of all those introns makes it possible for large numbers of them to sit around, having been turned on in the past to ensure their viability, but be turned on in a great wave due to environmental stress...

This is specious. It doesn't matter if a slight change in an existing gene causes a significant morphological change. If there are trillions of trillions of potential slight changes, why is a very small fraction of those potential changes retained?

You seem to be arguing for a logical switching on of certain genes under stress leading directly from one parent to one daughter branch. In fact however, if stress induces these changes (I think we agree on that) there is no reason to believe that only one line of new genes would be expressed, in fact its much more likely that many different newly expressed genes would show up in different offspring, in which selection would prove which are beneficial.

This would put you firmly in my camp.

Well in order for this to be true you'd need to insist that somehow the mechanisms for copying exons are different than copying introns so that no steering could occur without expression. While RNA strongly emphasizes copying exons (the SciAm 4/05 article shows we now know they copy introns too), the copying mechanisms are roughly the same and do the same "steering" to both types of genes..

Buff- I don't understand this argument. I don't know what introns do, but they are copied into mRNA just as exons are (by definition) or else they are not called introns. We do not know (that I am aware of) what the excised RNA fragments coded by introns do yet. I was not contending anything about the behovior of introns except that they are specificed in the genome, not "accidently" inserted and subsequently excised.

..You keep insisting that the new genes have to be constructed from whole cloth, while its clear that there is a highly structured process--that I think has been evolved, not preprogrammed--that can easily be explained by serial gradual mutation that is expressed in a punctuated fashion: And there's the basic crux of the matter: if there's any programmatic process in evidence, it must dominate the mutations. Guess what? We don't disagree that the programmatic functions built into the process *make it work*....

It seems to me that we have now expended a lot of text on element of agreement. You are agreeing that mutation alone could not drive speciation of a daughter species from a parent species without some sort of programming in th parent species to "steer" the mutation. Your disagreements are then only two fold:

1. You suggest that there was some number of mutations in the code that were the "fodder" for this steering mechanism, and
2. You are suggesting that the steering mechanism itself evolved, rather than being programmed in some progenitor

Do keep in mind that you are now about a millimeter from my position. The questions to ask yourself wither respect to item 1) above are: "What incremental value do mutations provide in this schema as an explanatory mechanism?" I believe the answer is "None." The second question is "What evidence is there for a role of mutation in speciation?" I believe the answer is "None."

The question to ask yourself with respect to the second question is "When would this steering mechanism had to have arrived in the phylogenetic tree to 'steer' the process (mutative or not)?" If we hypothesize that introns and transposons are part of the "steering" process, then they would have to have existed in prokaryotes to allow for these features arriving in single celled eukaryotes. If this biochemical service is in the prokaryote, you are inching toward exactly my position.

Do keep in mind that I offered this treatise only because you goaded me into it (although I did expect to have fun doing it- that certainly happened) . Did this discussion bend your thinking?

...You berated me in an earlier post about how unbelievably smart the self-correction mechanism would have to be, but your theory in fact requires a much more strict and complex mechanism...

I don't know which post that was, but clearly I am a weak communicator on this. I was attempting to use your assumptions to show that we agreed (as above). You are agreeing to a very complex mechanism to "steer" mutation. My argument was that your mechanism was not significantly simpler than mine. I just could not figure out why you assumed (by fiat) that mutations were a mandatory element without evidence to support it.

..it assumes that *all* future evolutionary changes are in there somewhere too--just saying they're "programmed" rather than pre-existing does not solve the information-theory related problem that they have to be in there somewhere.

Agreed. But I already thought that the prokaryote was so complicated anyway that making it multiple orders of magnitude worse did not really aggravate the problem.

...mutations...get away from the information theory problem...and its the 500-pound gorilla standing behind you.

Now we are getting really close. My point is that the gorilla was already 10 billion pounds. I made it 100 billion: it seems like sort of a "so what?" moment. Further, the mathematics for mutation are so weak, no matter how you account for them, that they don't really address any of the probabilistic problems associated with PE. I just opted for the larger gorilla since it is more consistent with the extant evidence. Further, this line of reasoning (i.e., preprogrammed speciation tendencies in the parent species) can be readily pursued experimentally. I suspect the pieces will be uncovered by accident anyway.

Remember you heard it here first, folks.

...Bottom line: if you're going to argue that they have no effect on major morphological/phenotypic changes, while at the same time arguing that they do make changes to genes, you need to explain how they are forever bypassed.

Gracious. This is the easiest one. Mutations are usually bad. When expressed, they result in death or a weakened species. They do not need to be bypassed. They terminate themselves. This is a MUCH better example of natural selection that Darwin's finches. I think that Darwin's finches (consistent with my hypothesis on speciation) were designed in the parent species. Ergo, they were not natural selection at all. Mutations are not specified in the parent species, and they select out. In my model, we would probably have to change the name of this process to "natural deselection" since the vast, vast, vast, vast majority of cases result in termination of the expressed gene.

[Biochemist]

Goodness. This gets the award for the "heaviest lifting" thread that I have participated in to date.

It is a pleasure to debate with you two- Buff and Bumab.

[Biochemist]

I just found this link:
http://id-www.ucsb.edu/fscf/library/...IONS/sub1.html

It shows some pretty good grahics relateded to the nature of stasis and the number of cases where the sudden start of phyla does not particularly well map to the notion of branching of the phylogenetic tree.

The final graphic (graphic I) is particularly instructive.

[Buffy]

Originally Posted by Biochemist

First, I was being hardly definitive. Second, I am not sure which is the stretch: believing that genes that preexisted suddenly launched new dramatic phenotypes (without evidence, and against probability) or presuming that the parent species had a prespecified tendency for change in the daughter species (with partial evidence).

If you cannot be definitive on this point, your entire argument breaks down! Your "second" point is again trying to say that if mutations have *any* effect in the process whatsoever, this means that their only application must be through instantaneous generation of new genes from scratch by random combination. It is amusing to me that you don't see that this black and white view of how things have to be is exactly why you end up sounding "definitive". You do not allow for shades of grey here. But we'll come back to that in a minute.

Originally Posted by Biochemist

I don't know what you mean here, Buff.

You are arguing massive changes (which I don't disagree on), but that these are entirely "new genes" when parent dna which no longer exists might well show some significant shifting and multiple branches. A key part of your argument seems to be that the sequence is:

1. Environmental Stress occurs
2. A SINGLE set of APROPRIATE genetic changes are caused that respond to the stress
3. A SINGLE new daughter species results

Evidence of this sequence could erroneously be inferred by comparing the dna of a current living example parent species to a currently living dauthghter species. It also implies that the dna changes are completely pre-programmed to take the effects based on specific stresses--which is a significant increase in the level of complexity in the programming your model requires in those cells! This is distinctly different from the model that I'm describing (which by the way, Gould appears to agree with):

1. Environmental Stress occurs
2. A LARGE NUMBER of DIFFERENT genetic changes are caused that respond to the stress
3. MANY changes in morphology result in MANY new daughter species, and under the stressful environmental conditions, only a few survive.
4. Environmental stresses can clearly span many generations, resulting in successive rounds of these changes, multiplying the number of changes that are expressed over the stress period

This introduces randomness as well, which I know you abhor, but we'll come back to that.

Originally Posted by Biochemist

It would be reasonable to assume that at least some of the (at least) several hundred new genes expressed in the 100 or so new phyla that presented at the beginning of the Cambrian period were "new" in that they were not previously phenotypically expressed. Even if some of their code was used previously, that does not really solve the mathmatical problem unless the vast majority (certainly higher than 99.9%) of each gene was used previously.

No, that is not at all a reasonable assumption. I know it is to you, but the sequence of events due to environmental stress I just described provide a significant mechanism for feedback. The only thing that makes your view reasonable and the other "unlikely" is that you do not allow for the likelyhood that a single stress will result in many morphological branches that are tested severely under the stressed environment. Not believing that such multiple branches would occur simulaneously--indeed that only one new branch occurs--defies logic, but its the only situation in which your math starts to work.

Originally Posted by Biochemist

It would be truly a stretch to suggest that significant portions of the functional, coded elements of the genome that have had no previous opportunity for phenotypical steering via natural selection, were "retained" and reused suddenly- unless the genome had a prescribed mechanism to do it.

First, in order to be functional, it probably was expressed, but would not have significant morphological changes without a certain minor changes made to it. These genes in fact seem to be extremely sensitive to such small changes! Moreover, due to their status as "functional" there's no reason to think that they would be excised!

Originally Posted by Biochemist

Yes, and the targets for copying are all specific. I am suggesting these are intended copies, not mutated copies.

So what? With all those copies, there's *lots* of opportunity for some of them to be mutated, and once mutated, copied slightly differently!

Originally Posted by Biochemist

I was trying to improve the probability for your position before I described that it was highly improbable under overly favorable circumstances.

Disingenouous! You were trying to mash my explanation into your warped view that mutation is all or nothing, which remains the weak point--indeed probably the disproof--of your arguement.

Originally Posted by Biochemist

If you assume that the majority of "junk" DNA is actually functional, your probabilistic case gets substantially worse.

No not at all. There is no inconsistency between believing that mutation causes useful changes and believing that the junk is mostly (or almost entirely) functional! It is only your insistence that "any mutation argument requires that all construction of genes be random" that creates any inconsistency, and that's your issue, which you still have yet to explain.

Originally Posted by Biochemist

There is not ANY evidence that mutation drives speciation. There is quite a bit of evideence that mutaion causes disease, but not speciation.

You keep saying this without explanation, but the surrounding arguments you make keep pointing to you believing that the words "mutation drives speciation" insists that unless its *all* mutation--without any role of the other mechanisms--then its in agreement with your argument that mutation plays *NO ROLE WHATSOEVER*. There is obviously no middle ground for you, which to me makes no logical sense.

Originally Posted by Biochemist

This is specious. It doesn't matter if a slight change in an existing gene causes a significant morphological change. If there are trillions of trillions of potential slight changes, why is a very small fraction of those potential changes retained?

Now you're backing up my argument above, that small changes can have an effect! And as for your 95%/99.99999% argument, have you ever noticed that these gizmos don't grow without bound? There is a limit to what's there, because DNA strands can obviously get "too big". The copying/steering mechanisms are designed to preserve the existing version--that's why they go for so long *without* changing over time--and so stuff has to get thrown out constantly. This is probably the explanation for why there are so many identical copies of functional genes in the introns: its a survival mechanism when stuff gets excised when the dna strands get too big. It takes evolution to grow them!

But lets cut to the chase:

Originally Posted by Biochemist

You are agreeing that mutation alone could not drive speciation of a daughter species from a parent species without some sort of programming in th parent species to "steer" the mutation. Your disagreements are then only two fold:

1. You suggest that there was some number of mutations in the code that were the "fodder" for this steering mechanism, and
2. You are suggesting that the steering mechanism itself evolved, rather than being programmed in some progenitor

Do keep in mind that you are now about a millimeter from my position. The questions to ask yourself wither respect to item 1) above are: "What incremental value do mutations provide in this schema as an explanatory mechanism?" I believe the answer is "None." The second question is "What evidence is there for a role of mutation in speciation?" I believe the answer is "None."

I'm only a millimeter from your position in your mind because of this continued insistence that "any mutation=completely random formation". That's your issue, not mine. I have consistently said throughout this thread that there are mechanisms for steering and reuse of functional genes. Now its fine that you believe that mutation has no effect, but you still have to deal with why they have no effect when you admit that they occur. You have avoided this so far by continuing to insist that any inclusion of mutations in the process requires completely random formation of elements, and I keep telling you it does not. You have not explained the reason for this particular position of yours, and there are plenty of examples from all sorts of complex systems--not just biological--that show that these random changes do affect the systems.

There is another interesting problem in your formulation that has occurred to me, that you'll need to explain: given that your model posits only single changes and they are all pre-programmed, then that pre-programming would have to know in advance the environmental changes that would occur in the future earth, and know when to express them. But lets take a look at the second issue, where we really diverge:

Originally Posted by Biochemist

The question to ask yourself with respect to the second question is "When would this steering mechanism had to have arrived in the phylogenetic tree to 'steer' the process (mutative or not)?" If we hypothesize that introns and transposons are part of the "steering" process, then they would have to have existed in prokaryotes to allow for these features arriving in single celled eukaryotes....You are agreeing to a very complex mechanism to "steer" mutation. My argument was that your mechanism was not significantly simpler than mine. I just could not figure out why you assumed (by fiat) that mutations were a mandatory element without evidence to support it....But I already thought that the prokaryote was so complicated anyway that making it multiple orders of magnitude worse did not really aggravate the problem.

Now, we both agree that there are two elements to the mechanism of change:

1. There are useful functional genes in the junk
2. There are steering mechanisms (a combination of the proteins, RNA, etc that do the copying)

You are insisting that

1. all of the information is pre-coded
2. the steering mechanism is more complicated than the data
3. the steering mechanism had to be part of the original programming

All of which I disagree with, and this is based on my own work with neural networks which have many of the same principles involved. First: in neural networks, the rules for steering can start out unbelievably simple. They can include mechanisms for feedback that not only affect the system being built but also *alter the feedback systems themselves*! All of this is unbelievably simple to code initially, and it becomes unbelieveably complex over time, *both* in the system and the meta-system (the steering mechanism). Most importantly one of the most important steps in making neural networks work is that they have a built-in mutation mechanism that says "well, we've always done this thing this way, lets flip a coin and try it a different way". This is the only way to find optimal solutions when there are isolated sub-optimal solutions.

As I say, this is unbelievably simple to code, BUT the only way you get to more complex systems is by *throwing in a random change*!!!!

This is the source of my argument that the explanation I have does indeed provide a much simpler solution. Yours requires an unexplainable appearance of a highly complex system (so'kay, you don't have to explain *why* it appeared, but there are testable hypotheses that I have drawn that seem to disprove it which you have not addressed), where as mine ends up not requiring much complexity for abiogenisis (which you may find philosophically troubling...sorry!).

One last point:

Originally Posted by Biochemist

Mutations are usually bad. When expressed, they result in death or a weakened species. They do not need to be bypassed. They terminate themselves....Mutations are not specified in the parent species, and they select out. In my model, we would probably have to change the name of this process to "natural deselection" since the vast, vast, vast, vast majority of cases result in termination of the expressed gene.

"Usually" does not equal always. So what if the bad ones are weeded out? If anything, this is proof that much of the change caused by mutations is not morphological, but rather microevolves species to better survive over time. There's lots of evidence that mutations provide the ability for those that have evolved protection mechanisms to survive. Is the fact that genetic resistance to desease in certain racial groups--provably caused by mutation--still not relevant to speciation? Given PE says speciation is a rare event, this point does not prove that they have no role. At all. Sorry!

Not sorry to have goaded you. Its been fun to debate this, ahem, odd theory of yours! I'm actually not convinced you haven't been running this entire thread with the entire contents of your tongue in your cheek....

Cheers,
Buffy

[Buffy]

Originally Posted by lindagarrette

Buffy, Your explanation makes good sense to me except for the above statement. How could inactive (junk) DNA combinations be "turned on" for some members of a species and not for others? What would cause the selection? Linda

Hi Linda!

There could be a large number of causes--that the boys and I have skipped over because we all seem to agree on this point of PE--hormones, overproduction of T-cells or other biological mechanisms, etc. Environmental stress does all kinds of bad things to us, including cell chemistry, which could have radical effects on gene copying, which could cause expression of inactive genes. This of necessity happens on an individual basis, and a key part of what we are arguing here is that I'm saying that these stresses cause a large number of changes to occur in different individuals, and that selection weeds out the useful ones. I *think* Bio is arguing that the genes are pre-programmed to express a specific inactive gene following pre-specified rules of how to respond to the specific environmental stress. In either case though, its still possible that both the parent and daughter branches could persist if they had slightly different environments (volcano forms a new mountain ridge and the environment on one side is very different than the environment on the other side, causing both branches to survive over time.

Cheers,
Buffy