Comments on Creation and Evolution Blog: Evolution, Chance, and Design (T.O. CB940)

Cool site. Thanks!!!Sport Watches Rubber LonginesH...

2006-06-26T16:20:00.000-07:00

Cool site. Thanks!!!
Sport Watches Rubber Longines
Hayes House Radio
Hp 1315 Photosmart
Fujifilm S5500 Webcam
L5r Card Search

Cool site. Thanks!!!Sport Watches Rubber LonginesH...

2006-06-26T16:19:00.000-07:00

Cool site. Thanks!!!
Sport Watches Rubber Longines
Hayes House Radio
Hp 1315 Photosmart
Fujifilm S5500 Webcam
L5r Card Search

Very good site. Thank you!!!Knot Of EternityTimber...

2006-06-26T14:08:00.000-07:00

Very good site. Thank you!!!
Knot Of Eternity
Timberland High Boots
Portable Mini Washer Washing Machine
Electric Freestanding Ovens
Open Arms Preschool

ubtvdsp rogt bljed bcktp hdnvml rewqguvn lorh byam...

2006-06-15T00:56:00.002-07:00

ubtvdsp rogt bljed bcktp hdnvml rewqguvn lorh byameqs zeir

mqlicng pcgo jlurgnvm hvfeons nfaswjh qugcrtnfl qn...

2006-06-15T00:56:00.001-07:00

mqlicng pcgo jlurgnvm hvfeons nfaswjh qugcrtnfl qnum http://www.hvsnpkgb.eylvqdnbo.com

http://all-too-common-dissent.blogspot.com/2006/01...

2006-01-23T05:30:00.000-08:00

http://all-too-common-dissent.blogspot.com/2006/01/haldanes-dilemma-another-creationist.html

"I guess that's why the editors of peer-reviewed j...

2006-01-19T16:55:00.000-08:00

"I guess that's why the editors of peer-reviewed journals have been telling him they won't publish because the issue is already well-known?"

This is quite incorrect. Editors told him they wouldn't publish on another issue because it was on material that had been published 25 years ago. It wasn't about Haldane's Dilemma.

"The fact is, you are basically limitted to 1,667 beneficial mutations. That's barely even enough to build a full gene, much less account for the beneficial differences between chimp and man."

Also quite incorrect. Since these are beneficial mutations, they are all gene modifying mutations. Walter has also acknowledged that in the same time, you could get 25,000 neutral expressed mutations. So, how many base pairs does this come out to be? I have gone a few rounds with Walter on this, using his own assumptions. Here is what Walter wrote:

"Evolutionists do not get to assign the 1,667 mutations any way they please, say, as "regulatory genes" or as "mutations with a large effect". Nature does not work that way. Rather, the preponderance of mutations will be of the ordinary kind, with a small effect. Let me illustrate the concept with crude figures: about 1500 mutations with an ordinary small effect, 100 more for re-positioning genes on chromosomes (inversions and so forth), 60 as gene duplications, and 7 mutations to regulatory genes that have a larger effect – for a total of 1,667."

The Human Genome Project estimates that the average gene consists of 3000 base pairs. Just taking Mr. ReMine’s assumed 60 gene duplications, this would provide for 180,000 new base pairs not present in the common ancestor, and this is just from a small fraction of the 1,667 beneficial mutations. It is obvious that 1,667 beneficial mutations and 25,000 neutral mutations could add up to a great deal of base pair differences.

The funny thing was that when I pointed this out to ReMine, he complained that I had misrepresented him. In dealing with him over the years, I have learned that anything he did not explicitly say - even if it is a given from his argument - amounts to misrepresentation if he doesn't like the outcome.

The Haldane argument is bunk, plain and simple. I have personally challenged ReMine to defend, and all he does is dodge and weave. See: http://www.baptistboard.com/cgi-bin/ultimatebb.cgi/topic/36/111/4.html?

"Um, no - I think that most of them are neutral. P...

2006-01-11T16:31:00.000-08:00

"Um, no - I think that most of them are neutral. Please slow down and re-think your statement."

There was a missing "not" in that statement. Thanks for pointing it out!

"In fact, they cannot be, for as I mentioned, any two humans differ by some several million nucleotides, therefore, it is impossible to determine how many substituions, especially neutral ones, are in fact fixed."

No it is not. It would be based on sequencing a wide range of humans. If the change exists in a wide sample of humans, it would be shown to be fixed. Most of the differences between chimps and humans are fixed. There are 35 million base substitution differences, as well as 5 million insertion/deletion events (totalling about 40 million nucleotides).

"You may have noticed that it is on a creationist site."

Yes. In fact, I wrote it. So what? Are creationists wrong by definition?

"ReMine the electrical engineer who has been misrepresenting the issue for years."

Really? I guess that's why the editors of peer-reviewed journals have been telling him they won't publish because the issue is already well-known?

"For example, he claims that 1667 fixed beneficial mutations are too few. But he does not know what traits the ancestor had!"

This is silly, because we know the number of mutational events between chimps and humans. Likewise, I would imagine that there would need to be at least that many changes just for going to obligate bipedalism.

"Further, in his terrible book, he implies that it would take more than 500,000 such changes if evolution were true. This is, of course, quite stupid - the human genome only has 25-30-,000 genes!"

Do you think genomes change an entire gene at a time? You should read Behe's peer-reviewed article about how long it takes just to change three _amino acids_. In addition, the regulatory elements are not included in gene counts, and are just as important if not more so. In addition, a better count would be of proteins, not genes, which then goes into the hundreds of thousands.

"ReMine acknowledges that neutral mutations can also contribute to phenotypic differences and that they accumulate at a higher rate than do beneficial mutations, not to mention single nucleotide polymorphisms."

This is true, but the rate of change would simply have to be astronomical, unlike any changes we've ever seen.

The fact is, you are basically limitted to 1,667 beneficial mutations. That's barely even enough to build a full gene, much less account for the beneficial differences between chimp and man.

That is correct, but do you honestly think that ou...

2006-01-11T08:47:00.000-08:00

That is correct, but do you honestly think that out of the millions of differences between chimps and humans, less than 1% of them are neutral?

Um, no - I think that most of them are neutral. Please slow down and re-think your statement.

I'm also pretty sure that the differences I am quoting from are fixed in the population, but I will have to double-check.

No, they are not. In fact, they cannot be, for as I mentioned, any two humans differ by some several million nucleotides, therefore, it is impossible to determine how many substituions, especially neutral ones, are in fact fixed.

For more information, see:

http://www.nwcreation.net/wiki/index.php?title=CB121

You may have noticed that it is on a creationist site. ReMine the electrical engineer who has been misrepresenting the issue for years.

For example, he claims that 1667 fixed beneficial mutations are too few. But he does not know what traits the ancestor had! Therefore, simple logic dictates that he cannot know how many are too few. I know this because I have asked him on many occasions and he simply ignores the question.

Further, in his terrible book, he implies that it would take more than 500,000 such changes if evolution were true. This is, of course, quite stupid - the human genome only has 25-30-,000 genes!

Further still, ReMine acknowledges that neutral mutations can also contribute to phenotypic differences and that they accumulate at a higher rate than do beneficial mutations, not to mention single nucleotide polymorphisms.

"Haldane's model dealt with fixed beneficial subst...

2006-01-10T10:05:00.000-08:00

"Haldane's model dealt with fixed beneficial substitutions, not all substitutions."

That is correct, but do you honestly think that out of the millions of differences between chimps and humans, less than 1% of them are neutral?

I'm also pretty sure that the differences I am quoting from are fixed in the population, but I will have to double-check.

For more information, see:

http://www.nwcreation.net/wiki/index.php?title=CB121

The number of mutational events that appear to hav...

2006-01-10T09:41:00.000-08:00

The number of mutational events that appear to have occurred between chimps and humans is in the millions. If you half that (since we are going from a common ancestor), you are still _far_ above 1,667.

You are completely unaware of what "Haldane's dilemma" actually indicates, that much is obvious.
If yoiu are going to refer to an issue like that, it seems reasonable that you should at least have an understanding of the basics.
Haldane's model dealt with fixed beneficial substitutions, not all substitutions.
As an example of how out of thye ballpark your take on Haldane's dilemma is, consider this:

Any two humans differ by millions of substitutions. According to your take, no two humans can be related...

The number of mutational events that appear to hav...

2006-01-10T05:43:00.000-08:00

Plus, you have the cost of selection, which limits...

2006-01-10T05:24:00.000-08:00

Plus, you have the cost of selection, which limits the number of mutational events (of any type, not just base pair) between man and ape's most recent ancestor as being about 1,667, which is far too few.

Please explain why, if that number is correct, it is 'far too few'.

I read over your blog, and i found it inquisitive,...

2006-01-07T23:02:00.000-08:00

I read over your blog, and i found it inquisitive, you may find My Blog interesting. My blog is just about my day to day life, as a park ranger. So please Click Here To Read My Blog

Excellent blog article. found your blog through CR...

2005-10-10T19:58:00.000-07:00

Excellent blog article. found your blog through CRSnet.

The origin of information is a brickwall that the evolutionists keep crashing into.

You're confusing the example. You said, "Whatever...

2005-09-18T17:06:00.000-07:00

You're confusing the example. You said, "Whatever phrase is closest to "methinks it is like a weasel" is kept and the rest are discarded, and it starts over from there. However, this does not get at the main problem in two separate places: (1) the intermediate forms make no sense whatsoever."

Actually, the intermediate forms do make sense as defined by the rules of the example. It's all about survival, my friend. You're operating with the assumption that evolution is working toward something (in this case an understandable phrase). This is a common misconception.

So, "Mathunks it as bike a veesel" makes more sense than "jhdfasjdgasjdgshd" because its construction allows it to survive and reproduce. It is more "fit" than a random jumble of letters. And as long as it is more fit --able to survive and reproduce-- it's existence makes "sense" (i.e. its construction is advantagous)

Wow.. I'm still digesting all of this information....

2005-08-16T20:43:00.000-07:00

Wow.. I'm still digesting all of this information. I'm very impressed by how you've collected all of this.

I'm not very "scientific" so this is almost too much for my finite mind to handle... =) But wow, this is very cool. Thanks!!!

"I presume that the first of those three component...

2005-08-16T04:35:00.000-07:00

"I presume that the first of those three components, the population of organisms, is what you refer to as the "small dataset", and the latter two collectively make up what you describe as the "very large base" of the program."

Not really. Perhaps if you were only referring to the genetic algorithm portion of a program that might be true to some extent, but then you wouldn't have a very useful program. Did you read the link I provided?

There are many problems with genetic algorithms within the context of modelling life. First of all, the functions for reproduction are not contained within the genetic algorithm at all. In an organism, pretty much everything is in the genome -- instructions for the organism to grow, develop, copy itself, etc. In genetic algorithms, none of these portions are subject to mutational effects, but in evolutionary theory these happen with the same regularity as everywhere else. Not only must a mutational change provide instructions for the adult, but also it has to be comprehensible to the zygote and intermediates as well. Also, the genome contains all of the instructions for the machinery itself. Second of all, the genetic algorithms cannot usually be subject to extinction. So, you can have the full freedom of the search space when nothing makes sense. Biologically, this would lead to extinction. For example, if state Y required passing through state X, but state X was an organismal dead-end, the biological species would go to extinction while the genetic algorithm would keep on chugging. Finally, the expressivity of the coding properties in genetic algorithms is (a) very low, and (b) has an inordinately large set of stable possibilities. In computer programming, expressivity is a direct relation to how chaotic the underlying programmed system is. For example, of all of the basic cellular automata studied by Stephen Wolfram, only one of them -- rule 110 -- was turing-equivalent. Rule 110 is also the most chaotic of them. Genetic algorithm programmers usually use non-chaotic "genetic" elements, which means that their programs will have both (a) an unusually high set of comprehensible sequences and (b) an unusually limitted set of possibilities. Life is even more chaotic than turing machines -- IIRC only about 1 in 10^11 proteins within a search space are reactive AT ALL, much less in a functionally useful way (please check my numbers -- I forget where this was referenced and may be off). Finally, organisms have multiple, relatively-independent subsystems, while genetic algorithms do not.

Anyway, I have to stop now, or I'll miss my bus. I'll try to finish up later, but I think most of my responses are simply going to be extensions of the above or a reference to the link I gave you earlier:

http://crevobits.blogspot.com/2005/08/genetic-algorithms.html

"The problem is that genetic algorithms do not wor...

2005-08-15T21:21:00.000-07:00

"The problem is that genetic algorithms do not work like evolutionists claim that animal genetics work."

I don't agree. Allow me to address your comments one at a time.

"Specifically, they do not allow random changes to programs. Instead, they allow random changes to a small dataset. What makes genetic algorithms (and any other non-deterministic algorithm, for that matter) work is that there is a very large base of the program that is immutable. The 'randomized changes' are only randomized within a specific subset of the program or data, and used to accomplish a specific, precoded purpose."

I don't think this conveys an accurate impression of how a genetic algorithm works. A genetic algorithm, in general terms, has three components: a pool of candidate programs (the "organisms"); a piece of code that causes those candidate programs to reproduce, mutate, and swap segments of code (I don't know of a standardized term for this, but let's call it the "evolver"); and a piece of code that evaluates candidate programs to determine how effective they are at solving the specified task (the "fitness function"). I presume that the first of those three components, the population of organisms, is what you refer to as the "small dataset", and the latter two collectively make up what you describe as the "very large base" of the program.

First of all, I take issue with your characterization of the relative sizes of these components. Neither the fitness function nor the evolver need be very large or complex - very often they can be quite simple indeed. Conversely, the population pool need not be small; depending on the parameters of the simulation, it can contain thousands or millions of members, and the complexity of any given individual depends on the complexity of the task it represents a solution to. It is entirely possible that the amount of code that mutates and evolves in the course of a GA far exceeds the amount of code that does not.

Secondly, I must object to your implicit claim that the immutability of the other two components is an issue. Of course the evolver and the fitness function are not subject to change. That is just what we would expect, and I would argue that that is analogous to organic evolution. Together, these two components set the rules for the simulation; they are not themselves part of the population, but instead create the "world" in which the population exists. In this respect, these components are analogous to the laws of physics, which define the parameters of an organism's environment, and which do not change.

So when you say this:

"...the changes can only occur within pre-coded bounds, and the program is in fact expecting there to be changes within certain parameters."

- that is the same as evolution in the wild. In organic evolution, too, changes can only occur within certain boundaries - those boundaries being within the genomes of living things. Mutations cannot occur independently of any living thing, nor can they change the physical laws defining the world in which those living things exist. Similarly, the "program" - life, the species, the genome, what have you - is likewise "expecting" change to occur, in the sense that genomes are structured so as to make it possible for evolution to happen.

Finally, you said:

"Remember -- genetic algorithms still take programmers to write. They have to be carefully crafted to work at all!"

I hope you realize that that is not an argument against evolution. At best, it is an argument for theistic evolution. And please do note that genetic algorithms can produce outcomes incorporating information that their programmers did not have.

In any case, the fact still remains that genetic algorithms, by combining random changes with a non-random process of selection, very often produce impressively complex, tightly efficient, and highly functional results. That is, of course, the essence of how the evolutionary process works in living things as well. Genetic algorithms are not an exact simulation of how evolution works in the wild, and it would be absurd to expect them to be; but they do show clearly that the evolutionary process constantly derided by creationists as incapable of producing true novelty, in reality, can and does work.

ebonmuse -The problem is that genetic algorithms d...

2005-08-14T16:23:00.000-07:00

ebonmuse -

The problem is that genetic algorithms do not work like evolutionists claim that animal genetics work.

Specifically, they do not allow random changes to programs. Instead, they allow random changes to a small dataset. What makes genetic algorithms (and any other non-deterministic algorithm, for that matter) work is that there is a very large base of the program that is immutable. The "randomized changes" are only randomized within a specific subset of the program or data, and used to accomplish a specific, precoded purpose.

If genetic algorithms worked like evolutionists think that organism genetics works, they would break themselves far before they produced anything useful.

So, while the changes are non-deterministic, to call them random would be incorrect, as the changes can only occur within pre-coded bounds, and the program is in fact expecting there to be changes within certain parameters.

Remember -- genetic algorithms still take programmers to write. They have to be carefully crafted to work at all!

Anyway, I've written about this a little more extensively here:

http://crevobits.blogspot.com/2005/08/genetic-algorithms.html

Hello,I must take issue with this comment:"If this...

2005-08-12T13:28:00.000-07:00

Hello,

I must take issue with this comment:

"If this sort of mechanism really worked, for example, it would make my job as a computer programmer a whole lot easier. To enhance a program, I could simply write another program to make random changes to it, and then test each one to see if it was a better program. I could remove the need to ever be creative! In fact, it could probably make changes I never thought of. Of course, programs don't work like this. The only thing you'll get in this case is a broken program."

Actually, there are programs that do work exactly like this, and produce highly complex, functional and efficient programs as an end product. They are usually called genetic algorithms, and they explicitly mimic evolution as a problem-solving technique: beginning with a population of daughter programs, they randomly mutate each one, evaluate the effects of these mutations to see if they are beneficial, then allow the beneficially mutated programs to "reproduce" and join the next generation. Genetic algorithms are already being used in industry to schedule flights at airports and assembly lines in factories, to evolve new antimicrobial compounds for use in cleansers, to increase the efficiency of engines and turbines, and more. There have even been instances where genetic algorithms, by making random changes to a program and then filtering those changes through a non-random process of selection, have produced results that outperform programs explicitly written by human beings to achieve the same goal. See:

http://www.talkorigins.org/faqs/genalg/genalg.html#examples

Your statement that "Of course, programs don't work like this" appears to be in conflict with the facts.

You cannot say that gene duplication always gets r...

2005-07-25T07:45:00.000-07:00

You cannot say that gene duplication always gets removed by further selection or whatever was your argument

Perhaps you should actually look at what my argument was. It was that duplicated genes are still subject to natural selection. The gene duplication theory hinges on duplicated genes being functionless, and thus being able to freely walk the search space without being selected upon until they are beneficial, and then being selected favorably.

What is found instead is that gene duplications are very functional, and that mutations in duplicated genes are still subject to selection, and therefore are not free to walk the search space as hypothesized in the gene-duplication model.

And, as I also pointed out, it wouldn't matter if it could because the chances of it finding a useful mutation even if it were given freedom from selection is minutely small. Plus, you have the cost of selection, which limits the number of mutational events (of any type, not just base pair) between man and ape's most recent ancestor as being about 1,667, which is far too few. See Haldane's Dilemma

You cannot say that gene duplication always gets r...

2005-07-25T07:29:00.000-07:00

You cannot say that gene duplication always gets removed by further selection or whatever was your argument. It might be what happens most of the time, but that doesn't mean it happens always.

You also seem to ignore the scale of these random-walk searches. Yes, it is glacially slow, but it isn't in a hurry either.

If you just want to believe in god, go ahead. But twisting facts to your liking and planting lies upon lies is inmoral.

"mutations can involve very large duplication of e...

2005-07-18T19:20:00.000-07:00

"mutations can involve very large duplication of existing genes. Once those genes are duplicated, they are free to mutate on their own."

This is nice in theory, but it's just not true. The fact is that duplicate genes are still functional, and when they suffer from mutations they are removed from the gene pool by purifying selection.

Gene duplication does not merely result in a second, redundant copy. In fact, there is some evidence that gene duplication is part of a Lamarckian-type genetic mechanism.

Your statement relies on two premises: (a) that duplicated genes are redundant -- this is false, and (b) that duplicated genes therefore don't have the same selective pressures and are therefore free to mutate without restraint -- also false. For more information on this, see:

http://www.pnas.org/cgi/content/abstract/0503922102v1

(despite its impressive title, the contents agree with the above)

In addition, let's pretend for a moment that your assumptions are true. If natural selection is not acting on a gene, that means that the _only_ things affecting them are mutations, which are more or less random. Therefore, you have an even larger problem than before, because you really do have to mutate through the entire search space in finding a gene. For information on why this is highly improbable, see this article from Protein Science on the ability to generate novel proteins from gene duplication:

http://www.proteinscience.org/cgi/content/abstract/ps.04802904v1

So, not only is the premise incorrect, but the conclusion is unlikely even given the premises.

"Furthermore, given classic information theory, any duplication increases information content."

That's true according to Shannon information theory, but that only measures information statistically, and ignores semantics. So, while that is useful, what is needed is an increase of semantic information in non-designed ways.

Here's an AiG article about the semantic extension to Shannon Information theory:

http://www.answersingenesis.org/tj/v10/i2/information.asp

For some quantitative looks at the semantic information content of genes, see:

http://www.trueorigin.org/spetner1.asp

"Educated readers will know, however, that DNA is nothing at all like computer code - comparisons of DNA to software are bound to be empty of meaning."

Like all analogies, there are flaws, but I think to say that it is "nothing at all" similar would be an overstatement, as they basically deal with the same types of constraints, though the coding for function is quite different.

This is incorrect, mutations can involve very larg...

2005-07-18T15:32:00.000-07:00

This is incorrect, mutations can involve very large duplication of existing genes. Once those genes are duplicated, they are free to mutate on their own. Furthermore, given classic information theory, any duplication increases information content.
Educated readers will know, however, that DNA is nothing at all like computer code - comparisons of DNA to software are bound to be empty of meaning.