Tuesday, April 11, 2006
Luck Favors the Prepared, Darling
Randomness comes in many varieties. There are almost as many definitions of randomness as there are people investigating it. However, the best definitions focus on unpredictability -- either total unpredictability or unpredictable with respect to certain events or viewers.
Darwinism uses randomness in the context of "random mutations". Darwinism is "random mutations" plus "natural selection". By "random", Darwinists do not think that mutations are not subject to physical constraints. Instead, what they mean is that the mutational process is not forward-looking, but instead a product of copying errors. Therefore, since it is the result of happenstance copying errors, there is no relation between the likelihood that a mutation will occur and its fitness value. From Berkeley's evolution website: "In this respect, mutations are random—whether a particular mutation happens or not is generally unrelated to how useful that mutation would be."
Now, the interesting this in, I do think that there are biological processes which exhibit some amounts of randomness, but without exhibitting any (or at least very much) Darwinistic randomness. Let's start out slowly with a few exercises.
First of all, let's say I have a set of numbers: [1 2]. If I pick one at random, I have a 50% chance of hitting either one. Now let's look at another set of numbers: . If I pick a number at random from this set, I have a 100% chance of getting the number 5. So, interestingly, the difference between randomness and determinism is the size of your set -- determinism simply being a random process with a set size of 1.
Now, let's say I have a list of 12 unordered numbers: [5 30 7 33 2 9 4 200 18 12 123 1]. Let's say I need to see if the set contains the number 12. What's the fastest way to do this? Answer: there's several, but an obvious one is to just scan from left to right until you find it. Now, let's assume that there are two of us searching. What's the fastest way now? Answer: divide up the numbers in some way, so we each only have to search 6 numbers. This search will on average be twice as fast as our last one. Just for fun, let's divide up the numbers in an alternating fashion - I get the odds, and my partner gets the evens. Now, let's add a third person -- now I get every 3 starting with 1, my partner starting with 2, and the new person starting with 3. This will be, on average, three times as fast. You get the picture. By increasing the number of searchers, we can dramatically increase the speed of the search. Now, let's do something different. Let's say that we don't know how many searchers are going to be available. What is the best way to divide the search up among an undetermined number of searchers? Well it just so happens that the best way to do it is to let each searcher choose numbers at random.
In computer science this is called non-deterministic programming. It's actually a rather good way to write parallel algorithms because you can do a search through a search space and it will scale easily just by adding computers.
And, in biology, that's exactly what you have. You have an environment, and a population trying to find the best genomic configuration for that environment. The individual organisms cannot "sense" the exact size of the population. Therefore, in attempting adaptational strategies, the best way to search the space is for each individual to search it randomly.
Wait... did I just say "search randomly"? Well, I did, but that's not quite the case. I could bore you with the probability arguments against how amazingly long it would take for a random walk in the genome to beneficially change just a few amino acids, but instead I'll leave that to Behe. I have personally made algorithmic objections to such random walks, which you can view at another website.
So, am I speaking out of both sides of my mouth? Random is the best way to search, but random won't get you anywhere? Don't we believe in a young earth?
The solution to this riddle is that the organism must reduce the search space to only reasonable choices. Then, from that reduced search space the best way to find the appropriate change is through a random walk of those specific options.
Often times you hear me talk about specific or semi-specific responses. This is why. Sometimes the organism can figure out from its current stressors exactly what it needs to do, and it can generate a specific response. Other times, however, the organism may not be able to determine the one best strategy. There may be tradeoffs that deal with possible future conditions. In cases such as these, the best strategy is to have a random walk of a constrained search space.
So, to be clear, these are non-random, informationally-directed genome changes. There is no theoretical restriction on the size of the leap, and it can go directly from stable point to stable point. However, when the organism has to decide between multiple pathways where it cannot fully decide the best direction, the best optimization strategy for the organism is to specify it at random.
The paper Chance Favors the Prepared Genome is the summary paper of an entire volume of search strategies that genomes use to pattern their search space to find beneficial changes. The volume, part of the Annals of the New York Academy of Sciences, is entitled Molecular Strategies in Biological Evolution. Just the table of contents makes for interesting reading.
The abstract of the lead paper is very telling of the volume's contents:
Genomes that generate diversity also are at an advantage to the extent that they can navigate efficiently through the space of possible sequence changes. a Biochemical systems that tend to increase the ratio of useful to destructive genetic change may harness preexisting information (horizontal gene transfer, DNA translocation and/or DNA duplication), focus the location, timing, and extent of genetic change, adjust the dynamic range of a gene's activity, and/or sample regulatory connections between sites distributed across the genome. Rejecting entirely random genetic variation as the substrate of genome evolution is not a refutation, but rather provides a deeper understanding, of the theory of natural selection of Darwin and Wallace. The fittest molecular strategies survive, along with descendants of the genomes that encode them.
There's a hat tip to Darwin at the end, but it is clear that they are completely removing the "random mutations" from the whole neo-Darwinian equation.
I have not read the volume, but the lead paper gives a good summary of the contents. Here are some of the strategies employed by genomes (or, more specifically, employed by God in genomes) mentioned in the paper (there are undoubtedly many more):
- Horizontal transfer of genetic material from one kind to another to be repurposed
- Gene duplications allows an organism to explore variation around an existing functional framework
- Modules within the genome that promote recombination at variationally-important spots (repetitive DNA is used to delineate these modules)
- These modules are segmented as "interchangeable parts" which can match and rearrange as functional units to produce new, unique, combinatorial units.
- Tandem repeats act as "tuning knobs" for the expression of certain genes.
- Genomes can participate in "coordinated multilocus changes" and regulate genome rearrangements.
- Predictable environmental challenges cause fairly standard responses, such as in the immune system.
- Unpredictable environmental challenges cause major genome rearrangements as organisms search the available search space for a response.
- Uptake of DNA from the environment occurs during times of nutritional stress.
- The ability of organisms to exchange genetic information indicates that they can learn from each other [note from me -- also indicates a functional reason for a universal genetic code]
- Transposon movement is unleaashed during times of stress.
- Organisms can induce mutations under stress by inducing double-stranded breaks and then repair them.
- Pathogens, when in a new host, increase the mutation rate of certain "contingency" loci which regulate pathogenicity, while not mutating more "core" functionality.
- RNA can explore new possibilities, and then when it finds "successes" it can reverse-transcribe itself into DNA.
- Hotspots of genetic change in RNA are non-random.
- RNA editting allows a genome to try out new sequences before incorporating them into the genome.
- The immune mutate specific regions of antibody genes to generate new binding sites quickly without affecting the structure of the antibody itself.
- Pathogens have recognition sites for inserting variation into surface proteins.
- Snails can rapidly generate toxins to respond to variations in predators, prey, and competitors.
- Genes likely contain information within them about which sites are available for site exploration.
- Oxytricia reorganizes its whole genome every generation.
It is important to note, however, that some of these mechanisms are based purely on comparative genomics, and include the assumption of common ancestry in their formulations. As Creationists, we reject common ancestry, and concern ourselves mostly with observed changes, or at most, changes within kinds. Given that the authors don't share our view, it is likely that we would disagree with some of the evidence for some of these mechanisms.
However, all of this together indicates that the way that genomes change is not via random sequence changes, but via highly constrained, directed walks. While some individual choice of steps may indeed be random (and I'm certainly not discounting random mutations in the degenerative sense), it seems that by and large they are the result of a search space that has been highly structured to produce beneficial, usable change. This allows for highly saltational changes, such as those proposed for the rapid post-flood diversification.
As Edna Mode would say, "Luck favors the prepared, darling.".
While many people simply feel that such structured search spaces are themselves the result of unstructured searches, I would point out the following two articles in defence of the design position:
Searching Large Spaces
From a Creationist standpoint, I think we also need to look at such adaptations in terms of an organisms purpose in the ecosystem. I think we might even be able to predict what kind of beneficial changes are available to an organism based on the organism's purposed role in the environment, and the current organismal functions available in the local environment [good lifelong research project right there[. This was reflected on a little bit in other posts.
Soon we're also going to look at ways that somatic tissue might be able to make changes in the germ line cells in response to stress. Life is looking more and more organized every day. What a wonderful creation! God has created us not only so we could survive, but that we could adapt quickly and beneficially and fill the earth, in accordance with His purpose.
This sort of simplistic preachiness is so hackneyed and cliche - can't you people come up with something a little less... silly?
Oh, God is so amazing, why my hemorrhoids are clear evidence of His Divine purpose in my Design!