|
Author
|
Topic: Information creation and transcendence
|
gedanken
Member
Member # 594
|
posted 30. January 2003 12:19
Warren said:
quote:
The claim or suggestion that ‘probability distributions’ can be used to explain dynamic environmental conditions in evolutionary theory is patently ridiculous. You can not claim that survival and reproduction depends on adaptation to average phenotype conditions. An organism adapted to an average environmental temperature will not necessarily survive the range of temperatures that may occur.
The misunderstanding of how science works embedded in this paragraph shows the irrelevance of the larger body of Warren’s post, so I will concentrate on this paragraph.
Neither I, nor did any scientist, advocate for “averaging” of parameters, and only making assumptions based on the average rather than the range of parameters. This would only work under certain cases of linear systems, in which the average of the result or the average of the input have a distributive property (e.g. commutivity of averaging operator). Clearly biological activity is non-linear, and not subject to this condition over large variations, and clearly scientists understand this in constructing their models.
Let’s examine a “genetic algorithm” (GA) as an analogous and partially descriptive model of biological evolution, with regard to: “The claim or suggestion that ‘probability distributions’ can be used to explain dynamic environmental conditions in evolutionary theory is patently ridiculous.” We are looking at the relevance of the statement “You can not claim that survival and reproduction depends on adaptation to average phenotype conditions” to that first sentence. And therefore the relevance of the rest of Warren’s post, and of many previous posts based on claim that dynamically time varying conditions can’t meaningfully be used in models of selection and evolution.
Of course the GA would simulate the distribution of conditions. When it discovered that the organism (member of the population) did not survive the extreme, that member would be eliminated. That elimination would affect the trend of overall population of simulated individuals with characteristics that can’t withstand the extreme.
So in this case we have a simple example of a model using a “probability distribution”, and it does not suffer from any difficulty as Warren suggests.
Where a more simple trend analysis could possibly be used in biology (and I am not speaking as an expert here in any sense) is that there would be a trend to eliminate organisms that could not stand the extreme of temperature. But even this elimination would be a statistical result. It would possibly show a decreasing trend of total population that have certain characteristics that can’t stand the extremes at some point in the dynamically varying environment. So the trend is involved in averaging the response, not the input conditions. (Why Warren should suggest averaging the input condition as something that scientists would consider solely is beyond me -- I certainly did not mean that when stating that statistical trend methods are important.)
With such a trend of not surviving the extreme temperature conditions of a particular time dynamically varying environment, we would have clear selection between organisms that have a structure to withstand and operate within those extremes, over those who do not. We have the linear average characteristics of sums of populations as the domain of the statistical trend projections, where we are within a linear domain of functional response. And thus the trend analysis is very useful in investigating the selection response of variations of organisms within a population, with regard to a distribution of dynamically time varying conditions.
Another point is very relevant. If we have a similar distribution of environmental conditions around a mean, like temperature, and then we note a change of the mean, we have clear information about the statistical nature of the extremes as well from that average. For the amount of time the environment will spend at a particular extreme temperature and what that extreme is will vary with the average. So in that sense, the average temperature is a perfectly valid input. With small variations, we may in fact find linear approximations to the responses to the time variation. With such linear responses, with small variations in the average input, we indeed have a case where we can commute the averaging operator with the (now linearized) response operator -- and in this case find useful case for studying simple average input to output response. In this case of small variation, we can clearly have a selection response trend dependent on an “average” of the “conditions” -- this is hardly ridiculous, it is simply a result of mathematical linearization techniques such as are expected from Tailor expansion of functions.
The appearance of these arguments seems to be a mutating random set of presentations of ways that science should and does not operate, without recognition of how science actually does operate and the consistency that is demanded to get published in mainstream peer-reviewed scientific journals. I could keep responding, but there are an infinite number of presentation of ways that science should not operate. It would be much more useful if Warren would present an objection to something that science actually published, and understanding how the conclusions were reached.
At this point I really can’t spend more time responding to this same issue, so I will only return to this thread to discuss issues that relate to the opening topic in a more direct manner. (I agree this is relevant -- but has been discussed at sufficient length now in my opinion) Thanks for the discussion.
[ 30. January 2003, 15:06: Message edited by: gedanken ]
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 30. January 2003 15:21
There are a number of posters on ISCID like gedanken who on the one hand seem to consider themselves experts on Darwinian evolutionary theory, and on the other hand make claims that exhibit a strange lack of understanding of scientific concepts. Darwinian and neo-Darwinian theories are descriptive theories and wannabe causal theories.
Basic forms of Darwinian theory suggest the environment impacting selection should be stable, but scientists know this is not reality. Basic Darwinian theory suggests that evolutionary change should be slow and gradual, but scientists known evolutionary change can be very rapid and discontinuous. Basic neo-Darwinian theory requires a functional genotype to phenotype map but no such map is known and the evidence clearly shows that if a functional map exists then it must be extremely complex. Scientists are fully aware of the contradictions between what is known and what is suggested by basic theory. Scientists, unlike some self proclaimed authorities, know these are issues which must, at some point in the future, be resolved before a predictive, causal theory can be formulated.
Science is not guilty of using intentionally misleading assumptions because scientists have been unable to formulate valid, rigorous, predictive, causal, real science theories of evolutionary change based on Darwinian concepts or any other concepts. While science can not be accused of making claims based on misleading assumptions, the same can not be said for many individuals making broad and unsupportable claims for existing evolutionary theories.
After claiming that I did not understand how statistical fluctuations could explain the apparent discrepancy between ‘expected static’ and ‘actual dynamic’ environmental conditions, he is now claiming I don’t understand because I point out that the statistical fluctuation explanation is ridiculous.
The issue being discussed here is concepts and definitions of information and information processing capacity. I proposed a mathematical definition for volume of information and volume of processing needed to generate information. Based on the proposed definition, which is based on fairly standard computer concepts, I suggested there are low volume of information models and high volume-rapid processing models.
Based on the definitions provided, and the available evidence, it seems unlikely that a ‘low volume of information model or theory (a simplistic Darwinian model)’ could produce a valid predictive model (without inappropriate manipulation of assumptions). This conclusion, may or may not be obvious depending on your knowledge of information processing concepts. I suspect that anyone knowledgeable in the area of Darwinian evolution already knows that simplistic Darwinian or neo-Darwinian concepts will not produce a valid predictive theory.
The point I have been attempting to make, is that ‘if’ as seems apparent, low volume information theories can not produce a valid predictive theory, then maybe we should seriously consider high volume-rapid processing-teleological models and theories. Seems like a fairly obvious, and not particularly controversial conclusion.
[ 30. January 2003, 15:25: Message edited by: warren_bergerson ]
IP: Logged
|
|
gedanken
Member
Member # 594
|
posted 30. January 2003 18:34
Warren, I am not an “expert” on evolution -- far from it. This is a thread on “information creation and transcendence”. My expertise is much greater in the subjects of information (as studied both in physics and engineering), artificial intelligence (including some knowledge of genetic algorithms), and I can speculate on religious issues of transcendence along with the rest of us. But your arguments as presented are easily shown to be wrong -- such that even a person so inexpert on the subjects of evolution as myself can still show the holes in your argument.
Your description of a “stable” environment did not allow for moment to moment variation. (In other words you pointed out that there were moment to moment variations, and then denied and continue to deny that a stable statistical variation in the dynamic time-varying environment can constitute a “stable” environment, even when that might be well supported at some level of time granularity by statistical methods.) But of course there is no such requirement for “stability” in evolution models, you are claiming this but not those who study evolution. Those who study evolution are only noting that there exist averages over distributions of characteristics -- and that there are averages or ensembles of behavior when examines over that aggregate of time, space, and other dimensions of such conditions. Others have very successfully pointed out the problems and you appear ignore those points.
Then you say:
quote:
After claiming that I did not understand how statistical fluctuations could explain the apparent discrepancy between ‘expected static’ and ‘actual dynamic’ environmental conditions, he is now claiming I don’t understand because I point out that the statistical fluctuation explanation is ridiculous.
The issue being discussed here is concepts and definitions of information and information processing capacity. I proposed a mathematical definition for volume of information and volume of processing needed to generate information. Based on the proposed definition, which is based on fairly standard computer concepts, I suggested there are low volume of information models and high volume-rapid processing models.
I did not claim that you did not “understand” how statistical fluctuations could explain … whatever…, I rather claim that your presentation is not based on such an understanding of how momentary dynamic variations can have aggregate effects that can be studied very reliably using statistical techniques. Your statements still refuse to acknowledge that scientists’ models can be accurate -- and give examples of moment to moment variations and claim that those make the statistical ensemble based models “misleading”.
I show how those models can be clearly based on sound statistical techniques -- and then you still claim that they are “misleading” without showing any sensible argument for that. (Go back to my point about linearliztion in cases like average temperature and show where there is any error in this. A “fitness function” can clearly be modeled as a function of average temperature, when the time distribution of extremes track that mean and the nonlinear consequences are of small enough variation that they can essentially be linearized. Scientists don’t fail to recognize that more complex models are needed to cover more extended ranges when they discuss a limited range example.)
You fail to acknowledge, for example, that an ensemble of dynamically time varying events (such as the minute to minute variations in the struggle of prey and predator) can have average overall effects on a population that are stable season after season. I’m not claiming that they are always stable, rather they can be unstable and quite variable -- and I think that evolutionary models account for this -- but Warren you don’t even present a recognition that such dynamically variable patterns could have stable average effects over specified populations and sub-populations.
Warren, you want others to be able to understand what you present, and claim to have your own models for information processing. Until you can present an understanding of how other models actually work, you will have a double strike disadvantage in trying to get others to understand what you are trying to say and how it differs from those other models.
And in your second paragraph I quoted, you seem to be claiming that you establish the topic of this thread. I’m sure that you establish the topic of what you want to talk about, but that does not change the accuracy of statements made by others (any more than “strawman” arguments change the accuracy of other models -- see below.)
Your earlier quote:
quote:
The claim or suggestion that ‘probability distributions’ can be used to explain dynamic environmental conditions in evolutionary theory is patently ridiculous. You can not claim that survival and reproduction depends on adaptation to average phenotype conditions. An organism adapted to an average environmental temperature will not necessarily survive the range of temperatures that may occur.
is clearly not relevant to any actual model of evolution, such as a “fitness function” because it does not reflect the fact that the scientists do understand that there is a relationship of average temperature and the ensemble of events that occur at the extremes of the temperature variation.
I have no objection to your trying to develop an alternate model of “information” processing, that might indeed become very interesting. Just your claims of errors being made by others who have models seem always to make claims of things that those models don’t say, and then try to show that those mistaken versions would not produce reasonable results. (I think this is called making a “strawman argument”.) To do so is not going to get others to understand your own models. To demonstrate a new model, one must often compare it to an older model and show the differences. Presenting the older model as it actually is understood would be helpful in showing such differences -- so you must first present the older model accurately as it is actually understood by the originators.
If you can extend what is known, this would be most interesting and valuable. But to disagree with the accuracy of what is already presented by scientists, you have to show errors either in logic of what was presented or divergence with observation. An example is how quantum mechanics and relativity show that classical mechanics is in error. But classical mechanics is not in error in all circumstances, rather it does not cover all cases. So QM and relativity must have a “correspondence” principle such that they “reduce” to the older classical mechanics in the cases that are well supported by observation. Warren, I suggest that you find the “correspondence principle” to what is already shown to be well supported by observation. (That would require accurately representing that theory.)
Some relevant internet sites (not the important original literature, but just some pointers to issues that relate):
Punctuated Equlibria (at Talk Origins) Relates to points about “change can be very rapid and discontinuous”
Introduction to Evolutionary Biology relates to basic “predictions” of evolutionary models.
The Evolution of Improved Fitness Though this is primarily written as a response to creationism it also includes notes about information theory and “fitness” in very general terms. This is useful in understanding how momentary variations can still result in an overall “fitness” of an organism measured in likely lifetime or rate of reproduction or the like. These are the bases of evolutionary models -- and presence of different time grains of dynamic change don’t invalidate the longer time grains.
Good luck with your quest.
[ 30. January 2003, 21:02: Message edited by: gedanken ]
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 31. January 2003 02:46
quote:
Based on what is currently known of genetics, the number of possible developmental pathways for a given set of genes is very, very large. Only a very, very small portion of those pathways will produce a full grown organism capable of reproduction.
If by "possible" you mean "every gene could flicker at random on and off", then I agree.
However, essentially all of those developmental pathways are completely spurious to consider, since they are physically prevented due to coordinated regulation of genes. In a real organism, you can't get any arbitrary pattern of expression of genes 1-30000 at any instant in time. In fact, this is the whole point behind gene chip analysis of expression profiles used to find coordinated regulation of expression patterns. (E.g. based on the overexpression of a heat shock protein, loss of a developmentally important gene, or whatever.)
quote:
Again based on current knowledge, it can be demonstrated that the probability of a full grown living organism developing from an initial egg cell is far less than 1 in 10^1000 or even 1 in 10^10,000.
If this is back to your expression-profile calculation before, your probability calculation is assuming a uniform distribution over protein expression space, which as I mentioned above, does not at all correspond to reality.
quote:
If we look at cells making up a multi-cellular organism at any point in time, we known there are a very large number of different types of cells all developed from a single set of genes and single initial fertilized egg cell. The 'adaptive set of cell types' at any point in time is a very, very small subset of the set of possible cell types at a point in time.
True, and this small subset is specifed for by means of various developmental cues, ranging from polarity-establishing genes in the fertilized oocyte through to notch-delta signalling enforcing alternative devleopmental pathways for neighboring cells. We certainly don't know all the developmental pathways, but we know enough to start to get a picture of how the system works, and from what we've seen, it works by a series of causal relationships.
The cell types are not chosen uniformly at random. That really would be a hopeless way to conduct development.
quote:
In terms of genetics, we know that the assumption that genetic material does not change or 'evolve' during the organisms lifetime is false. In fact, we know that the state or form or regulatory genes are continually changing (evolving) during the organisms lifetime.
This is a bad use of terminology. Firstly, "genetic material" usually refers to the nucleic acid sequence in the genome--and that (with very rare exceptions, such as VDJ recombination in immune system cells--and this occurs only in somatic tissues and thus is not passed on) does not change during the organism's lifetime. Secondly, the state of regulation of genes over time is not evolution in the sense of Darwinian evolution. Yes, expression levels of genes change over the lifespan of an organism, but you don't need multiple rounds of evolution to get the change right then, because the capability to change in response to stimuli, time, etc. is a consequence of the way the organism is constructed.
quote:
If you look at the number of possible paths defined by possible changes in regulatory genes during an organisms lifetime, and if we look at the portion of paths that are likely to result in the development of living reproducing organism, we know there is not and can not be a causal relationship between genes and fully developed organisms.
Developmental paths are not chosen uniformly at random, nor are regulatory gene expression patterns. So it doesn't really matter that most paths result in an organism with zero fitness: those paths are overwhelmingly improbable.
Why can't this type of regulation be sufficient on its own?
You perhaps even hint that in can be when you say
quote:
Basic neo-Darwinian theory requires a functional genotype to phenotype map but no such map is known and the evidence clearly shows that if a functional map exists then it must be extremely complex.
Yeah, it looks complex enough. What do you think developmental biolgists spend their time trying to figure out? It's not straightforward--but that doesn't mean it isn't there.
quote:
Back to your original point. Not only is the claim of a causal or functional relationship between genotype and phenotype false, it is clearly known to be false.
I'm afraid that at this point I have to conclude that you are not sufficiently informed about developmental processes and gene regulation in order to make a case for your point. You have some imaginative ideas, but they're rather under-supported right now.
If you have a case to make, it must be much more biologically sophisticated in order to even be relevant. Specifically, you need to address the point of genes that are regulated together (and the consequent reduction in the space of plausible expression profiles); the point of low-information designs having high-information-processing capability; and the question of experimental results showing what looks like a causal relationship in many developmental systems (hox genes and segment identity; notch-delta signalling and sheath/socket/neuron differentiation; and on and on and on)--i.e. you must explain what about these experiments leaves room for non-causality.
-----------------
I have a bottle of water on my desk here. I'd estimate that it contains on the order of 10^24 molecules. Nothing can travel farther than 186,000 miles in a second, so each second the molecules have to stay inside a volume of about 10^26 cubic meters. However, for me to recognize it as a water bottle, and for it to maintain structural integrity, each molecule has to stay within about a 0.1mm of its original position, or 10^(-12) cubic meters. Thus each molecule could occupy any of 10^38 positions every second, so the probability of the entire water bottle staying in position is
1/(10^24)^(10^38) = 1 in 1,000,000,000,000,000,000,000,000^100,000,000,000,000,000,000,000,000,000,000,000,000
so obviously there must be an immense amount of active teleological design behind my water bottle staying in one piece on my desk--the probability of it doing so on its own is infinitesimally small.
See how silly the result of the calculation is, when you don't take relevant physics into account?
Likewise with how silly a calculation can be if you don't take relevant biology into account.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 31. January 2003 11:06
Rex,
Quote: However, essentially all of those developmental pathways are completely spurious to consider, since they are physically prevented due to coordinated regulation of genes
You are missing the point. Let me try one more time. To begin, we both agree that in most instances the developing organism normally is successful at selecting or finding one of the adaptive developmental pathways. This observation supports the existence of a descriptive causal relationship. This does not mean that mal-adaptive pathways are spurious. It means that the organism generally has the information required to take an adaptive pathway. We can agree that the normal success in finding an adaptive pathway suggests the existence of a causal developmental process. The question is what type of causal processes are involved. Basic neo-Darwinian theory, or ‘low information’ models would suggest that the information to find the adaptive pathway is present in the fertilized egg cell. In mathematical terms, the egg cell contains ‘a set of assembly instructions or a genotype to phenotype map’. The alternative explanation, I propose, is that the egg cell contains ‘teleological information processing capabilities’ which make it possible for the organism to generate much or most of the information needed to find an adaptive pathway.
The fact that developmental processes are normally successful and the developing knowledge of the physical mechanisms responsible for developmental processes are potentially compatible with both the causal map explanation and my teleological explanation. The initial question being raised is ‘Can development be explained entirely by the causal map explanation’ or ‘Does development depend on some combination of causal maps and some type of information generating process?’ Only if we answer the first question will it be worth addressing the two types of causal processes.
I suggest there is evidence that 1)the causal map explanation alone is not mathematically possible and 2)there is positive evidence that the teleological processing explanation does fit the available data.
To understand the ‘positive evidence’ supporting the teleological causal explanation, you need a reasonably extensive understanding of the mathematical concepts of scientific teleological causation and the physical mechanisms responsible for producing teleological processing. Because of the complexity of the positive arguments, I have been focusing on arguments/evidence that the causal map explanations(by itself) can not work. There are two of these arguments.
First, as I have shown, a set of genes does not map to a single phenotype but to literally billions of improbable adaptive states or phenotypes which are essential to survival. By itself, this does not disprove the causal map explanation. However, if the genotype to phenotype mapping is one to billions, then change operating only via variation in phenotype and natural selection would be extremely slow. We know from selective breeding experiments, that large numbers of related changes in adaptive states can occur very rapidly. This is not logically or mathematically consistent with the type of map suggested by the existence of billions of adaptive states. Build a model if you don’t believe me.
The second argument involves the required complexity of the causal map. A single predefined developmental map, a sequential listing of assembly instructions for developing a complex organisms, would involve a very large volume of information. But a single map would not be adequate to explain observed developmental processes. The organism does not in practice fail when a single assembly instruction fails. Instead, it can and does adjust or adapt to disruptions. To adjust for this with predefined maps would require a separate instruction for each contingency. The amount of information needed for predefined assembly maps grows rapidly if you are to account for possible contingencies. There is, I believe, lots of engineering evidence on the ineffectiveness or lack of robustness of predefined assembly instructions for complex assembly. The benefits of using teleological feedback in complex assembly instructions is also well documented.
The available evidence, I suggest, clearly shows that causal map explanation by itself is not adequate. Developmental processes involve a combination of ‘information stored in the fertilized egg cell’ and ‘information extracted from the environment during the developmental process’. The functioning fully developed organism could not develop entirely from the information contained in genetic material. If you reject this conclusion, then you need to be able to demonstrate that 1)the organism can develop from predefined assembly instructions, and 2) you need to be able to demonstrate that these predefined assembly instructions can evolve. I don’t believe anyone actually believes that developmental processes are actually controlled by evolving causal maps.
If you accept the obvious that developmental processes involve BOTH inherited information AND information generated within the organisms life time, then it makes sense to try to measure the relative volumes of inherited and generated information. To accomplish this you need a mathematically precise definition of information that allows you to quantify and compare volumes of information at different points in time, and to quantify and compare the volume of processing associated with different volumes of information. The definitions of information and information processing I proposed make it possible and practical to quantify information and information processing.
The problem is not, as you suggest, lack of knowledge of genetics and developmental processes. The problem is that biologists have failed to develop techniques to quantify information and information processing. Because of an inability to quantify information, the essential component of biological design processes, biologists are unable to properly interpret the available evidence. It seems ridiculous when you think of it, but biologists can not quantify the differences in information or complexity between a single cell and a full grown functioning organisms. Biologists have no effective means of demonstrating or disproving whether or not evolutionary change can be produced by a ‘1 cycle per lifetime process’.
The failure to quantify information is obviously as much a problem for ID as it is for EV. Because of an inability to quantify information, Dembski suggests that 10^150 is a level of complexity which can not evolve by natural processes (10^150 is a relatively trivial level of biological complexity). Mike Gene (sorry Mike), proposes to study front loading, but is unable to differentiate a front loading process that impacts information processing by a factor of 10 from front loading that impacts biological information processing by a factor of 10^1000.
Gedanken at the beginning of this thread was right in pointing out the importance of appropriate definitions of information and information processing. However, I don’t think he fully recognized how seriously the lack of an appropriate definition impacts evolutionary biology (and AI ).
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 01. February 2003 11:52
SELF ASSEMBLY
Biological systems have the capacity for complex self assembly. Rex claims that biologists and geneticists belief that this self assembly is the result of ‘causal instructions’ or a ‘mathematical causal map’ stored in genetic material. If Rex is in accurately describing the current ‘scientific’ position and not some common academic misinterpretation, then ‘self assembly’ provides the ideal for comparing and evaluating ‘ID the science’(design science) and ‘genetics the science’.
As I said earlier, I doubt that scientists actually support the claim of a genotype to phenotype map or that ‘genes determine the organism’ is causal relationship. In ‘real genetics’, I suspect, ‘genes determine the organism’ is a convenient simplifying assumption or myth offered as a euphemism for ‘we have no way of explaining self assembly’. I am sure there must be someone out there who is willing to state whether ‘gene to organism causal maps’ are a scientific reality or a scientific euphemism.
However, whether or not genetics has a scientific/causal explanation for self assembly, design science does. Design science, (at least my version) claims that the DNA ‘recipe’ for an organism can not be converted into an operating organism without an ‘intelligent design process’. The transformation from single cell to fully developed organism, I suggest, could not occur without the existence of an intelligent, information generating process capable of generating the information needed to select the appropriate assembly process or pathway.
Self assembly is an ideal subject matter for comparing ‘conventional scientific approaches’ and ‘design science approaches’ because we can analyze self assembly in the laboratory, and we can model self assembly mathematically. We know that self assembly occurs naturally. Snow flakes are an example of self assembly. We can create different artificial self assembly processes and measure the amount of information and information processing required to make these assembly processes robust under realistic conditions. We can compare different types of assembly processes not only for volume of information and volume of information processing, but also for degree of robustness.
We can readily segregate and analyze small components of self-assembly. The dubious argument that ‘its too complicated to actually model and test’ clearly does not apply to the analysis of self assembly. It should be noted that just as you can actually model, analyze, and simulate self assembly, you can also analyze processes that would change or evolve self assembly.
It is probably not as widely recognized, but it is fairly easy to define ‘dynamic and teleological’ assembly instructions. It is also fairly easy to demonstrate mathematically how dynamic and teleological assembly instructions evolve (and they do not and can not evolve using random variation and natural selection.).
Finally, it is useful to note that the analysis of self assembly takes the analysis of ID and biological design outside the realm and/or area of expertise of biologists and geneticists. The knowledge of biological processes needed, at least initially, to analyze self assembly is fairly minimal. Extensive training in biology provides little advantage in such analysis. The required knowledge of mathematical modeling, however, is well beyond the levels of the vast majority of biologists and geneticists. The analysis of self assembly not only offers an opportunity to demonstrate the value of the ID approach, but it also offers the opportunity to separate this analysis from the dubious mathematical and scientific standards currently used in biology.
IP: Logged
|
|
Moderator
Administrator
Member # 1
|
posted 01. February 2003 22:53
This thread has one more week of life left in it, at the most. I'll be shutting it down soon, so feel free to make any closing comments.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 02. February 2003 01:50
Hm, I could have sworn I posted a reply. Are the forums still eating posts? They really oughtn't any longer.
quote:
A set of genes does not map to a single phenotype but to literally billions of improbable adaptive states or phenotypes which are essential to survival. . . .We know from selective breeding experiments, that large numbers of related changes in adaptive states can occur very rapidly. This is not logically or mathematically consistent with the type of map suggested by the existence of billions of adaptive states. Build a model if you don’t believe me.
Here's a model. It is wrong, because it ignores the relevant biology. However, much as your calculation sets an upper bound, this sets a lower bound.
Suppose the billions of different possibilities are listed and labelled with an integer. This integer can be represented using 30-40 bits. Selective breeding that keeps, say, 5% of the progeny at each level can generate about 4 bits of information per generation. So in 8-10 generations, we can, by selective breeding, pick any of these billions of different possibilities.
Maybe you would like to modify your scenario to exclude this model.
quote:
I am sure there must be someone out there who is willing to state whether ‘gene to organism causal maps’ are a scientific reality or a scientific euphemism.
I think I have stated this a few times before, but we do not have definitive proof for the entire process, but we have isolated what appear to be causal maps when studying aspects of development. In the absence of evidence to the contrary, the assumption is that these causal maps (which may be impacted by environmental factors) are sufficient to specify development.
quote:
We can readily segregate and analyze small components of self-assembly. The dubious argument that ‘its too complicated to actually model and test’ clearly does not apply to the analysis of self assembly.
And the "self-assembly" of biological subsystems are modeled and tested, albeit less frequently than I'd like. Look through, for example, the Journal of Theoretical Biology. There's some really nice work on self-assembly in cAMP waves (spirals) in Dictyostelium, too.
IP: Logged
|
|
warren_bergerson
Member
Member # 262
|
posted 02. February 2003 11:27
Rex,
The basic issue here is defining and quantifying biological information and biological information generating capacity. In the absence of appropriate mathematical definitions, I suggest, it is not possible to perform rigorous scientific analysis.
You and I appear to agree that developmental processes or self assembly are an appropriate subject for evaluating my claims. I suggest that biological self assembly is ‘a complex biological design process’ which can not and does not occur without the presence of a very powerful ‘intelligent design process’ with a vast information generating capacity’. You advocate the more traditional view that self assembly is largely the product of a ‘genotype to phenotype map’ based on information stored in genetic materials.
I think you would agree that since we have 1)the ability to model both ‘dynamic and teleological self assembly’ and ‘genotype to phenotype type self assembly’, and 2)we have the ability to artificially simulate both types of self assembly, then, we should have the ability to determine experimentally which type of self assembly model more accurately fits biological developmental processes. We can not, however, resolve the issue by exchanging claims of what has and has not been done.
In order to resolve the issue, we need to review some basic concepts such as 1)assembly instruction, 2)fixed or permanent assembly instruction, 3)dynamic and teleological assembly instruction (programmable instruction), and 4)robustness. I will post my definitions/interpretations of these concepts on a new ‘Simulating Self Assembly’ thread.
IP: Logged
|
|
|
Printer-friendly view of this topic