Annoying creationists

[articulett]

Interesting, anyway.

If we have a printout of that Word file depicting only the underlying zeroes and ones, is that 'information'? I tend to say, no: that's 'data'.

Whatever you call it, it's a "code"--just like DNA is a code. It contains information--even when there is no-one there to translate it...even if th computer is off--the code is there. DNA is similar to that code. Whether you print that code in it's raw form or in it's "more readable form"--it's still a code. It can be translated into useful or useless information just as DNA can--if it's useful information in has a better chance at getting passed on than if it's useless or harmful code.

 

[articulett]

As I previously pointed out, there is more data in living things than is to be found in their DNA sequence.
I see no point in regurgitating the entirety of my web site on this thread. I suggest you read the index page and then either read on by your own interests or stop if you have none.

Would you say that there is more data on a computer then can be found in the computer code on that computer? What you're saying is open to interpretation.

If you imagine some other kind of data, how exactly would you measure for it? What would happen if someone was missing this extra "data"? I think you've made a leap here that is incorrect, and you are using a semantic game to claim there is this "extra data" that can't be accounted for by our current understanding of DNA, energy input, and so forth.

 

[kleinman]

Annoying Creationists

Random mutations and natural selection has always been the slogan for the theory of evolution. Random point mutations and natural selection has been the cornerstone for the theory.

This may at one time had been the case, but I don't know a single biologist who would hold to this.

I am very interested in hearing how the other mechanisms of mutation, gene and/or chromosome duplication and recombination with/without error will replace random point mutations and natural selection as the driving force for evolution.

The way Dr Schneider has written about ev, he has given very high weight to random point mutations and natural selection as the driving force for evolution and he is a molecular biologist. The editors of Nucleic Acids Research published these results as well in 2001. So if biologists no longer think that random point mutations and natural selection is the cornerstone for the theory of evolution, this is a recent development.

Ev demonstrates how profoundly slow this process is even with a very stringent selection process. I can not think of and have not heard a plausible explanation from evolutionists how other forms of mutations such as gene duplications, frame shifts, etc. would overcome this mathematical deficiency that ev demonstrates in the theory of evolution.

Here is where we disagree. This model shows that point mutations alone within a ceratin functional window can form new binding sites. Any argument from a time line is erroneous. There exists too many different types of stimulii that can speed up the process.

I have always been very careful to claim that ev shows that evolution is profoundly slow when realistic genome lengths and mutation rates are used. I understand that there are input parameters (short genomes with high mutation rates) that show that binding sites can evolve mathematically very rapidly. If that is the functional window you are going to choose, you will have to do this without any physical proof. It is not clear to me that there are other mechanisms that can speed up this process. Recombination without errors can give rapid divergence of a species such as what Darwin observed with finch beaks or what we see with dog breeding but there is no proof that this can lead to a new species.

Consider the very example you raised of Flu vaccinations. The concept of immunization appeared before Darwin was born. Attempts at immunization for smallpox were done before 1800. The concept of immunization was not drawn from the theory of evolution.

look into the idea of original sin and immunizations. Flu shots can (and have been shown) known to enhance a person's likelyhood of contracting the disease. Genetic drift in the antigen sites reduce the effectiveness of the antibodies generated from a vaccine. If the drift is huge and the antibody raised is completely ineffective, the body can use a rapid antibody generation method to fend off the infection. If they drifted only slightly, the body trys to rework the existing antibodies to target the virus. this is a slower process and typically results in a enhanced likelyhood of catching the disease.

I don’t know what you are talking about with original sin and vaccinations. In addition, if flu shots enhance a person’s likelihood of contracting disease, the whole field of preventative medicine is turned on its head.

There are limitations on genetic drift. Changes in the sequence of a gene can not be so great that the protein coded for no longer functions properly. Natural selection will select out that organism. When the CDC and other organizations try to determine the particular flu vaccine that will be used in any given year, they try to anticipate the strain. They usually look to what strains are showing up in the orient and develop vaccines against these strains. You aren’t advocating against influenza immunization? In 1920 there was a world wide epidemic that killed 20,000,000 people. That is more than the number of people killed in WW1. The last time I checked the CDC data on influenza deaths was during the SARS scare. At that time, influenza was killing over 30,000 people/year in the U.S. alone when SARS was killing maybe a couple hundred.

The game is, we can never really know which virus strain will propogate and how much genetic drift it will exhibit. Some change more than others. Our ability at understanding this viral evolution allows us to better select a range of viruses to immunize against. Also, it has suggested that a 4 strain (instead of the 3 used)selection spaced ideally across the epitope space would greatly minimize the likelyhood of "original sin" from occuring. This is directly a feature of evolution that has been modeled and effectively used in handling medical care.

I believe the variation of strains of influenza viruses represents a microevolutionary process. I don’t believe there are an unlimited number of strains of influenza viruses.

What analysis of the data are you talking about? Why is Paul running a series with a wider binding site? The only thing Paul has been doing is retracting statements which support my contentions. I’m wondering when Paul is going to throw the entire ev model out.

I guess we've been reading different threads. I've seen ev work quite well within the Rcapacity bounds.

You should know that one of the reasons I annoy Paul is that I like to quote him. When Paul runs his series with the wider site width (larger Rcapacity), I’m sure that he is going to have data that shows that the evolution of binding sites on a 100k genome with a mutation rate of 10^-6 is going to take around 200,000,000 generations as he previously extrapolated and has since retracted. I would agree that ev is working quite well—at proving evolution by random point mutations and natural selection is mathematically impossible, it is too slow. I’ll make sure I quote you when Paul generates the data.

I am considering co-opting Paul’s concept of Rcapacity. What is preventing me at this time is that I can’t tell whether this is a peculiar characteristic of Dr Schneider’s model or whether it represents something which would occur with a real genome. I suspect that in reality, natural selection doesn’t have the precision that Dr Schneider has given his model with his weight matrix and that Rcapacity is a reflection of this effect.

I have a feeling you are right. But it also provides another look at an interesting fact. Binding sites don't evolve in a meaningless genome. In the model, every site can mutate and has no seperate positive negative effect. Such a system would be unlikely to evolve. However, the genomes didn't evolve that way. They seemed to have grown with time. At least current theory doesn't suggest it and this model agrees with that. We don't start with a informationless background genome.

I disagree, the first living things had to evolve from an information-less background genome but that is not the point of the ev model. Ev only simulates the evolution of binding sites, the rest of the genome only contributes if a binding site happens to be identified on this region and this is considered an error and selects against that particular organism.

Let’s do a remake of the Scopes-monkey trial, you can play the monkey.

That's fine. You can be...the defendant.

This would give a lot more publicity to this issue than your dumb ass marketing plan. I think I would like this. What are you going to charge me with? Is it my calling you greedy, or is it the lazy part that you find so uncivil? Perhaps it is your inattentiveness? I take back the Dilbert part; at least Dilbert has some technical skills.

Maybe we can get this on Judge Judy.

If you want to engage in a scientific discussion on this topic, you have to come up with a workable definition that people can agree on. I think I see the distinction you are making between the words “data” and “information”. For example data could be the measured state of a system while the degree of order of the data would represent the information content.

Unfortunately, one would wait a long time for people to agree on definitions. The distinction I make between data and information is standard in school level IT but it tends to be ignored by experts in the same field. In the context of standard biochemistry, the genetic data would be the sequence of bases in a gene, while the biochemical activity of the gene's product (most often the enzymatic activity of a protein) would be information, the interpretation of that data.

Have you looked at Dr Schneider’s ev program? Do you believe that Dr Schneider has applied IT in an appropriate manner?

If selection is as powerful as to be able to organize atoms and molecules to create the genetic molecules and proteins necessary for life and then evolve these complex molecules to the life forms we observe today you should be able to mathematically model such a profound force. Dr Schneider’s use of a binding protein (weight matrix) which imposes a very stringent selection process yet is still profoundly slow in evolving binding sites.

Selection is quite logical but I do not think one can necessarily model it mathematically, one may be able to do so. Computer simulations are generally more powerful than mathematical modelling. My work is a metabolism first theory and my concern is with the selection of sun induced chemical oscillations to produce metabolic pathways. It may be techically possible to model that process but I think such simulations are beyond my capacities. I therefore confine myself to the general chemical picture.

I view computers as the tool which enables the computation of values in a mathematical model. Mathematical models which yield closed form algebraic solutions are usually so limited that it is very difficult to model a very realistic system. Unless the theory of evolution can be mathematically modeled to account for the needed genetic changes to evolve one species to the next, the theory will remain a soft science.

I don’t know how heating and cooling of the primordial soup would induce chemical oscillations to produce metabolic pathways. Ultimately, metabolic pathways are energy conversion processes, for example it can be photosynthesis for the production of glucose or the Krebs cycle for the breakdown of glucose for ATP. What type of energy conversion process to you see occurring in the primordial soup from these chemical oscillations?

These questions were address to joobz. I know you understand why you are doing a series with a wider site width. I’ll stick my neck out here and guess that the wider site width will not give a significant reduction in the number of generations for convergence. The series you are doing now will show that your extrapolation of 200,000,000 generations for the 100k genome with a 10^-6 mutation rate and a population of 1meg was reasonably accurate.

Here is my data so far:

population 32
binding sites 8
weight width 9
site width 10
1 mutation / 512 bases

genome size, generations

1024, 8000
2048, 20000
4096, 34000
8192, 37000
16384, 76000
32768, 272000
65536, 392000
128000, 1000000 (est.)

If you take your estimated value of 1,000,000 generations and assume a mutation rate of 10^-6 which is about 2000 times slower than the mutation rate you are using for this series, you get 2 billion generations for convergence. Increasing the population to a million would reduce this number by an order of magnitude giving 200,000,000 generations. This is the same value that you extrapolated previously. One improvement for your argument is that you are evolving 160 loci instead of 96 loci in about the same number of generations.

In the context of standard biochemistry, the genetic data would be the sequence of bases in a gene, while the biochemical activity of the gene's product (most often the enzymatic activity of a protein) would be information, the interpretation of that data.

So there's no information in DNA? If not, then how does the information leak into the proteins?

Hey Kleinman, we can stop having this conversation. There's no information in DNA.

And give up my status as the annoying creationist, forget it. You aren’t getting out of this that easily.

 

[joobz]

I apologize I don't have time to address all your comments. I'm swamped with work.

I have always been very careful to claim that ev shows that evolution is profoundly slow when realistic genome lengths and mutation rates are used.

this is not true. Paul, Dr. Adequate, and others have demonstrated why your values are not "realistic".

I don’t know what you are talking about with original sin and vaccinations.

Deem, MW, Lee, HY, "Sequence space localization in the immune system response to vaccination and disease" Phys Rev Letter, 2003, 8, 91:068101

In addition, if flu shots enhance a person’s likelihood of contracting disease, the whole field of preventative medicine is turned on its head.

Not at all. Although, it does demonstrate that healthy individuals shouldn't get the flu vaccine. Plain and simple.

 

[kjkent1]

This would give a lot more publicity to this issue than your dumb ass marketing plan. I think I would like this. What are you going to charge me with? Is it my calling you greedy, or is it the lazy part that you find so uncivil? Perhaps it is your inattentiveness? I take back the Dilbert part; at least Dilbert has some technical skills.

Maybe we can get this on Judge Judy.

<shrug> in response to your lame attempts at juvenile baiting.

On the substantive issue, I find your attempts to discredit various other selection methods as not capable of increasing the speed of evolution unpersuasive. But, that's the beauty of the Internet. Critics never actually have to prove an affirmative case. They can survive by simply criticizing.

The blunt fact, which you cannot avoid is that there are a myriad of life forms in existence, and if they are not the product of some natural process, then they are the product of magic, because there is no third alternative.

So, it's really up to you, whether you continue to bait others while wearing a cape and wand, or whether you actually contribute to the advancement of science.

Frankly, I couldn't care less.

 

[John Hewitt]

@1131 by John Hewitt
OK, that's helpful. If I recall your definition of "rank", a rank3 influence at rank1 sounds to me as being simply a nomenclature change. Intelligence being a sexual selection factor amongst humans is not a novel idea, for example.

I am not quite sure what you mean by nomenclature change. Rank1 evolution is essentially biological evolution, though some data omitted from genetics is included, such as the data that defines the sequence of genes on a chromosome. Yes, sexual selection is Darwin, The Descent of Man, and he does argue that the brain is a consequence of sexual selection - and I agree. Coevolution between biological and social evolution should transfer traits from social to biological by the Baldwin effect.

Could you point me at something in your website that might compare or contrast your approach to the more common treatments of similar ideas?

Multilevel selection is Sloan Wilson, Jablonska is another, but apart from myself I know of nobody who analyses the different levels from a foundation in data.

Edit: Oh, and you should consider Plotkin's book, "the nature of knowledge" and Popper's "three worlds" interpretation.

 

[John Hewitt]

"Data on its own has no meaning, only when interpreted by some kind of data processing system does it take on meaning and become information."

I don't like it. Sounds like an informal definition of information is being used. I wonder if the person was trying to make a distinction between uncertainty and information, but got confused.

Anyway, it doesn't matter too much.

~~ Paul

You are free to dislike it but it is the position taken by a good number of people in IT. That being so, it does matter.

Consider, for example, a number sequence 0120145673

You might interpret that data as you social security number or as my phone number or as any of several other interpretations. Knowledge would arise from the selection of one of those interpretations.

 

[Paul C. Anagnostopoulos]

If you take your estimated value of 1,000,000 generations and assume a mutation rate of 10^-6 which is about 2000 times slower than the mutation rate you are using for this series, you get 2 billion generations for convergence. Increasing the population to a million would reduce this number by an order of magnitude giving 200,000,000 generations. This is the same value that you extrapolated previously. One improvement for your argument is that you are evolving 160 loci instead of 96 loci in about the same number of generations.

Why do you think increasing the population from 32 to 1 million would reduce the generations by an order of magnitude? Our population experiments show that it varies as p^-.25.

~~ Paul

 

[Paul C. Anagnostopoulos]

You are free to dislike it but it is the position taken by a good number of people in IT. That being so, it does matter.

Consider, for example, a number sequence 0120145673

You might interpret that data as you social security number or as my phone number or as any of several other interpretations. Knowledge would arise from the selection of one of those interpretations.

Certainly we can interpret that digit sequence in any number of ways. Perhaps what you're getting at is that, if we don't know anything about the origin of the digit sequence, then speaking of information is misleading. Then certainly we should just refer to it as data.

Still, if we have a set of digit sequences such as the one above, we can calculate the information measure of the set according to Shannon. If that measure is positive, doesn't it make sense to talk of the information in the set? If it does not, then you need another name for the information measure.

And here's an interesting thing: If that digit sequence is entirely random, but we interpret it as a social security number, then we have gone from data to knowledge without information. I guess that's how belief systems work.

~~ Paul

 

[Paul C. Anagnostopoulos]

Now that the 128K run is finished, I'll republish the results:

Here is my data so far:

population 32
binding sites 8
weight width 9
site width 10
1 mutation / 512 bases

genome size, generations

1024, 8000
2048, 20000
4096, 34000
8192, 37000
16384, 76000
32768, 272000
65536, 392000
128000, 991000

(fits -14946 + 7.55G) So at 1 mutation per million bases, that's about 2 billion generations. If it varies as population^-.25, as our population experiments show, then at a tiny population of 1 million it would take about 63 million generations.

~~ Paul

 

[John Hewitt]

Certainly we can interpret that digit sequence in any number of ways. Perhaps what you're getting at is that, if we don't know anything about the origin of the digit sequence, then speaking of information is misleading. Then certainly we should just refer to it as data.

Still, if we have a set of digit sequences such as the one above, we can calculate the information measure of the set according to Shannon. If that measure is positive, doesn't it make sense to talk of the information in the set? If it does not, then you need another name for the information measure.

And here's an interesting thing: If that digit sequence is entirely random, but we interpret it as a social security number, then we have gone from data to knowledge without information. I guess that's how belief systems work.

~~ Paul

Yes, knowing the origin of a data string is one of the things that would give you and interpretative context for it. In general, information is data to given interpretative context or subject to an interpretative process.

It is worth remembering that, in Shannon's time, the words information and data were treated synonymously. Hence, he wasn't attaching any significance to his choice of "information" rather than "data." It is only the subsequent change in usage that causes this problem.

"If that digit sequence is entirely random, but we interpret it as a social security number," then I think we would be applying a wrong interpetation. We would then be in danger of following an incorrect knowledge system and, possibly, even of having trouble with the tax man.

 

[Paul C. Anagnostopoulos]

It is worth remembering that, in Shannon's time, the words information and data were treated synonymously. Hence, he wasn't attaching any significance to his choice of "information" rather than "data." It is only the subsequent change in usage that causes this problem.

But this shows that the change in usage is window dressing. Shannon used the primary term uncertainty. Perhaps we would have spoken of "data measure" instead of "information measure," but what difference does it make? There is still information in the data! It makes no sense to say that there is data in the data.

"If that digit sequence is entirely random, but we interpret it as a social security number," then I think we would be applying a wrong interpetation. We would then be in danger of following an incorrect knowledge system and, possibly, even of having trouble with the tax man.

Let's say I have some mysterious data sitting in front of me and I embark on a forensic analysis to determine what the data represents. I start with no a priori interpretation. How do I get anywhere, if there is no information in the data?

I think the distinction between information and knowledge may make some sense, but the distinction between data and information is funky. The distinction appears to be the result of a bunch of IT professionals needing a separation between data and the processing of data.

~~ Paul

 

[John Hewitt]

But this shows that the change in usage is window dressing. Shannon used the primary term uncertainty. Perhaps we would have spoken of "data measure" instead of "information measure," but what difference does it make? There is still information in the data! It makes no sense to say that there is data in the data.

Let's say I have some mysterious data sitting in front of me and I embark on a forensic analysis to determine what the data represents. I start with no a priori interpretation. How do I get anywhere, if there is no information in the data?

I think the distinction between information and knowledge may make some sense, but the distinction between data and information is funky. The distinction appears to be the result of a bunch of IT professionals needing a separation between data and the processing of data.
~~ Paul

I am sure that the distinction between information and data does arise from a professional need, but that does not make it invalid, just the reverse in fact. Moreover, it is not true, by this definition, that the information is in the data. Data is a matter of pattern and the statistical likelihood of the pattern existing by chance. Information is the resultant of data and some interpretative context of process.
As I said before, the same data can, when acted upon by different interpretative modes, give rise to different of information - telephone number or social security number. Conversely, different pieces of data can give rise to the same interpretation, "The sky is blue" or "le ciel est bleu." The same meaning arises from different interpretative modes.
The same situation arises in biology. As is widely reported, a DNA sequence in phage phiX174 codes for three different proteins in three different reading frames. Thus we have the same data producing three different interpetations. Conversely, in many species, we have two or more different variants of genes that produce proteins with the same enzymatic activity. Two pieces of data producing the same interpetation.

 

[RecoveringYuppy]

...but apart from myself I know of nobody who analyses the different levels from a foundation in data.

Can you point me at some part of your website, hopefully the most direct example you can think of, that will demonstrate this difference?

By "nomenclature change" I mean a change of terms used to identify things. Such as you using "rank1" to refer to something that most other people would refer to as "biological evolution", at least according to what you just wrote to me.

 

[kleinman]

Annoying Creationists

In addition, if flu shots enhance a person’s likelihood of contracting disease, the whole field of preventative medicine is turned on its head.

Not at all. Although, it does demonstrate that healthy individuals shouldn't get the flu vaccine. Plain and simple.

Perhaps you believe that healthy individuals shouldn’t get the polio, measles, mumps, rubella, tetanus, chicken pox, diphtheria, whooping cough, tetanus, etc. vaccines as well.

If you take your estimated value of 1,000,000 generations and assume a mutation rate of 10^-6 which is about 2000 times slower than the mutation rate you are using for this series, you get 2 billion generations for convergence. Increasing the population to a million would reduce this number by an order of magnitude giving 200,000,000 generations. This is the same value that you extrapolated previously. One improvement for your argument is that you are evolving 160 loci instead of 96 loci in about the same number of generations.

Why do you think increasing the population from 32 to 1 million would reduce the generations by an order of magnitude? Our population experiments show that it varies as p^-.25.

Every population series we have run shows a very rapid drop off in the rate of decline of generations of convergence as population is increased. This is one of the many extrapolations you make that I will not argue with you when you retract it.

Now that the 128K run is finished, I'll republish the results:

Here is my data so far:

population 32
binding sites 8
weight width 9
site width 10
1 mutation / 512 bases

genome size, generations

1024, 8000
2048, 20000
4096, 34000
8192, 37000
16384, 76000
32768, 272000
65536, 392000
128000, 991000

(fits -14946 + 7.55G) So at 1 mutation per million bases, that's about 2 billion generations. If it varies as population^-.25, as our population experiments show, then at a tiny population of 1 million it would take about 63 million generations.

We are in agreement with the extrapolation for 2 billion generations however; I think your extrapolation of 63 million generations is probably a little low. I think the number is probably closer to 200 million generations. Still, your numbers are not that far different than mine.

Your previous extrapolation of 200 million generations to evolve 16 binding sites, each 6 bases wide with a mutation rate of 10^-6 and a population of 1 million does not look so idiotic.

Ev reveals a mathematical problem in your theory of evolution.

 

[fishbob]

Perhaps you believe that healthy individuals shouldn’t get the polio, measles, mumps, rubella, tetanus, chicken pox, diphtheria, whooping cough, tetanus, etc. vaccines as well.

Comments such as this are very useful to readers of this discussion in clarifying your reasoning and logic capabilities. In case there was any doubt. Thanks.

 

[John Hewitt]

Can you point me at some part of your website, hopefully the most direct example you can think of, that will demonstrate this difference?

By "nomenclature change" I mean a change of terms used to identify things. Such as you using "rank1" to refer to something that most other people would refer to as "biological evolution", at least according to what you just wrote to me.

This is discussed in the bioepistemic evolution page.
The correspondences are
Rank0 evolution = prebiotic evolution
Rank1 evolution = biological evolution
Rank2 evolution = evolution within organs of the body using input from externally derived data and selected within the receiver. Example, evolution within the brain (see Edelman and other studies into the brain as a Darwinian machine)
Rank3 evolution = Evolution of social knowledge
Rank4 evolution = One example is professional knowledge. (Compare with structure of Popper's logic and the evolution of scientific knowledge.)

 

[kleinman]

Annoying Creationists

Perhaps you believe that healthy individuals shouldn’t get the polio, measles, mumps, rubella, tetanus, chicken pox, diphtheria, whooping cough, tetanus, etc. vaccines as well.

Comments such as this are very useful to readers of this discussion in clarifying your reasoning and logic capabilities. In case there was any doubt. Thanks.

I can’t tell whether this is a sarcastic remark or not. Immunization is probably the most cost effective part of the medical system. When joobz suggests that healthy individuals don’t need flu vaccination, I doubt he makes this recommendation with any experience caring for someone with influenza. This is a very debilitating disease. I have seen a patient suffocate to death from influenza induced adult respiratory distress syndrome. There is a small risk and cost with obtaining an influenza vaccination but this is a gamble I tell people is worth taking.

 

[Paul C. Anagnostopoulos]

As I said before, the same data can, when acted upon by different interpretative modes, give rise to different of information - telephone number or social security number. Conversely, different pieces of data can give rise to the same interpretation, "The sky is blue" or "le ciel est bleu." The same meaning arises from different interpretative modes.

We have two different definitions of information here. If IT professionals want to use information to mean "knowledge that arises when we select a way to interpret data," then they need to come up with another term for "Shannon information, "Kolmogorov information," etc. Perhaps they have to use "information measure" or something.

~~ Paul

 

[Paul C. Anagnostopoulos]

Your previous extrapolation of 200 million generations to evolve 16 binding sites, each 6 bases wide with a mutation rate of 10^-6 and a population of 1 million does not look so idiotic.

Yes, it still does, because that model will not converge at the 100K genome you were talking about. If you're going to object to extrapolation in general, surely you should object to extrapolation in an impossible scenario, right?

Ev reveals a mathematical problem in your theory of evolution.

Not due to that particular experiment, it doesn't. The question now is: How do the generations vary with population in my latest model? I've selected the 4K genome size and I'm running experiments with increasing populations. Then I'll do the same with the 8K genome so we can see the difference.

~~ Paul