Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Conundrum Regarding Evolution Probability

  1. Jun 30, 2010 #1


    User Avatar
    Gold Member

    I'm trying to get some real numbers for the probability of evolution occurring by random mutations, but cannot find anything helpful in the literature--- except that a single base-pair genome mutation occurs once in every 10exp-6 DNA replications.

    So I'm trying to use this meager information to calculate the probability for the evolution of the 2.9 billion base-pair human genome. My approximations for the entire genome produce horrid numbers. To simply the calculations, I'm trying to figure out the evolution probability for a single 900 base-pair gene.

    Since there are only four possible base-pairs, the problem appears simple. (To keep it simple, I assume that the probability of a new base-pair addition needed to build a gene is the same as that for a mutation.) The sequence of base-pairs is critical.

    Let's define a "correct" base-pair as one that continues the sequence for a particular gene.

    When a new base-pair appears, the probability that it will be "correct" is simply 0.25. It looks like the probability for a complete, correctly sequenced gene is then, 0.25exp900, which I calculated as 1.4x10exp-50.

    This seems unreasonably low. Given that there are about 23,000 functional genes in the human genome, the probability for all of them coming together right is a downright ugly (1.4x10exp-50)exp23000. That number chokes my calculator, might keep a Cray minicomputer busy for a few milliseconds, and seems too tiny to even be worth figuring.

    I'm only trying for reasonable approximations, but these numbers are absurd. In their face, random mutations cannot have produced a single small gene. (By comparison, there are only 10exp80 atoms in the known universe.) And I've yet to attempt to use the base-pair mutation probability to calculate the number of generations required to produce a human body.

    Why bother, when the anticipated result looks like we need more generations than the number of microseconds elapsed since the Big Bang?

    Where have I gone wrong?
    Last edited: Jun 30, 2010
  2. jcsd
  3. Jul 1, 2010 #2
    i think you ought to consider a large population as the starting point and not a single organism. so, you start with a large variation and the organisms with a sequence that is closer to the "correct sequence" will survive with a higher probability in each time step (maybe it is appropriate to consider evolution as a markov process).
    however, it is unjustified to assume that there is only one correct sequence, as variation of sequences corresponding to the same protein are large in different species, often proteins with the same function share less than 50% sequence homology.
    lastly, i think you have to consider other pathways of DNA evolution, e.g. sexual recombination etc.
  4. Jul 1, 2010 #3


    User Avatar
    Gold Member

    Thank you for your reply.

    There are a few problems with your first suggestion.
    1.) I'm trying to keep things simple and clear. Assuming a large population muddies the water.

    2.) There does not exist sufficient data for even a professional evolutionist to establish boundary conditions for a solution of the large-population problem.

    3.) If there were, it would not help. I'm no good at statistical math.

    4.) Surely, Darwinists with a particular interest in performing these calculations, have tried and failed. Else their Nobel prizes would be well known, for they'd have driven a railroad spike through the heart of Creationist ideology.

    5.) A given base-pair mutation is surely Markovian if it is also random. But the next step is selection, and that would appear to be memory dependent.

    That is useful information. Makes it tough to simplify the calculations though.

    I'll need to research that further. Can you offer a starting source? (It'll have to be via computer or books I can find, since I have no access to a university library.)

    Might not the sequence variations produce proteins with the same amino acid sequence, but, folded differently? That would cause them to function differently. Do the analytical techniques used to identify the proteins differentiate between various folding patterns?

    Oh, I did! Believe me when I say that I'm trying to reduce the problem to a simple, therefore soluble form. My understanding of sexual recombination is that it determines which variation of a particular gene is inherited, but would not change the inherited gene unless a random mutation occurred during fertilization or perhaps early development. That would be subject to the 10exp-6 mutation rate.

    I do not know enough to evaluate the "etc." category. Is there anything in it that would simplify the calculation?
  5. Jul 1, 2010 #4


    User Avatar
    Science Advisor
    Homework Helper

  6. Jul 1, 2010 #5
    i understand your desire to keep things as simple as possible. what i'm proposing is that perhaps certain things that are not considered simple have to be introduced for this highly complex process.
    i suggested to check the markov property with regards to transitions from one population state to the next in a discrete (verly large) time step, under the assumption that the environment stays constant. after some finite number of time steps, you should get to a stable state with some distribution of organisms in the "correct sequence" states. the selection information is contained in the transition probabilities, so i see no reason why there should be any memory of states other than the current one.
    the source that i learned about sequence homology in homologous proteins is biochemistry by stryer et al. i'm sorry that i can't provide you with page numbers, i don't have the book at my disposal right now. but if i remember correctly, there is one chapter dedicated to evolution although a lot of useful information is also scattered in the other chapters. if you are really interested in life sciences, i think it is a good point to start.
    proteins with the same amino acid might not originate from the same DNA sequence, since the map from genetic code triplets to amino acid via tRNA&tRNA transferase is not injective. thats one point that i forgot to consider. what i was trying to say was that there are proteins for the same task that do not share the same primary sequence of amino acids. what makes proteins work are usually specific patterns of amino acids, e.g. containing charged groups or reactive groups that determine their functionality (or structure for cis residues), along with secondary structure elements e.g. alpha helices etc. that determine the geometrical configuration of the protein (which is important e.g. in a key-lock type of situation, note that many amino acids sequences can lead to the same secondary strcuture element). most amino acid replacements do not concern the functional groups nor the secondary elements, so that the state of a fully functional protein is not unique in terms of amino acids sequence.

    my knowledge on sexual recombination is sketchy at best, but I think you are right in that the DNA sequence is not actually changed but that complete libraries of the human genome are recombined meaning that every such organism has two possibly slighty different versions of the same gene (emphasizing even more that there does not exist one correct sequence). many mechanisms exist that cut and paste large chunks of DNA between different locations of the same of different chromosome (this is usually how multimers, complexes of many domains of the same kind were created). there is an unbelievable variety of other mechanism that might play a role in evolution: alternative splicing, epigenetics, post-transcriptional chemical modification of bases.

    to summarize: i think a simple model of evolution does not exist. you have to be prepared to deal with complexity if you really want to begin to understand evolution. as a last word: i really have no good grasp of this topic, everything i said should be taken with a grain of salt. please do check my statements in textbooks (e.g. stryer)
  7. Jul 1, 2010 #6
    im my opinion, you are underestimating the scientific community. there must be some efforts to model a large population, googling turned up many results, also wrt to regarding evolution as a markov process. i haven't acutally looked at this, but you can if you want to http://www.google.com/url?sa=t&source=web&cd=9&ved=0CFQQFjAI&url=http%3A%2F%2Fbustamantelab.cb.bscb.cornell.edu%2Fdocs%2FBustamante_05.pdf&ei=IistTPWxF6eIOK2v8MUB&usg=AFQjCNFvn5odnP2yv4ho_MEkChHg3wGnqA&sig2=5J63-HXc4e8B1B6AkNxW8A" [Broken]
    keywords: population evolution markov chain

    i don't think creationist will ever be impressed by our reasoning.
    Last edited by a moderator: May 4, 2017
  8. Jul 1, 2010 #7
    Actually genetic recombination happens quite often during meiosis where it is called http://en.wikipedia.org/wiki/Chromosomal_crossover" [Broken]. During this process homologous regions of one maternal and one paternal chromosome overlap and exchange certain parts of DNA. So in fact sexual reproduction is not just imparting a certain subsets of ones maternal chromosomes and a complimentary subset of ones paternal chromosomes but also some chromosomes bearing information from both the maternal and the paternal one.
    Last edited by a moderator: May 4, 2017
  9. Jul 2, 2010 #8


    User Avatar
    Gold Member

    Now I understand why, shortly after my offspring asked a "Why" question with respect to some aspect of the world, their eyes would soon glaze over!

    I certainly appreciate your trying to help. But lacking doctorates in either biology or probability theory, there's not a darn thing I can do with that information. Sorry, but I really did try to make it clear what I'm looking for---

    Just a quick and dirty way to APPROXIMATE the probability of a gene coming into existence thanks to the single force of random chance.

    I did not request a simple model of evolution. I've read Darwin and a dozen aftermarket books on the subject, including Behe. I requested nothing more than a way to estimate a simple probability.

    Astronomers are masters at estimates, because that's all they can do. They don't always get it right, but at least they put out the numbers. And they are limited by their inability to perform experiments on the objects of their interest-- all they can do is observe.

    I am beginning to fear that evolutionary biology is not actually a science.

    You are right about creationists, but I'm beginning to understand why. My OP was part of a bet between me and a very bright creationist, who predicted that I would NEVER get a straight answer to my query. She compared your reply to a presidential obfuscation, and I have to admit that she's right. And I've no doubt that you gave this your best shot.

    Before posting anything on this subject I tried to locate probability estimates in the literature and internet with zero success. So far, this attempt has been a failure on many fronts.

    I've already lost the first of several bets, and now must take her to an expensive dinner at a chick restaurant with snooty kids pretending at being waiters, instead of a steak house with friendly waitresses. During this dinner she will know better than to talk to me about her almighty God, but she's already made some good points against Darwinism, and this obfuscation is not helping my case.

    Nonetheless, thanks for trying.
  10. Jul 2, 2010 #9


    User Avatar
    Gold Member

    Good information, thank you!

    Does a "homologous region" mean a gene of the same type, or something else entirely?

    Suppose we have a gene in a male chromosome which codes for a specific protein. Will its female counterpart merge with that gene in meiosis in such a manner as to create an entirely new version of the gene which codes for a different or modified protein? If so, how often does this occur?

    You obviously know more about this than I ever will, so I have another question. I've been trying to figure out how DNA codes for structure, but cannot find any information on that subject. Can you point me in a useful direction, preferably via the internet?
    Last edited by a moderator: May 4, 2017
  11. Jul 2, 2010 #10
    many processes are inherently complex and if studied with simple approximations yield conundrums as you described in your first post. Please have a look at Levinthal's paradox of protein folding. An approximation qualitatively similar to your own, leads to the paradox that a single protein folding into its native structure would require a time that exceeds to age of the universe.

    if you are unwilling to dig (much) deeper into this subject, i'm afraid you are going to lose your bet. Enjoy the dinner with the lady.
  12. Jul 2, 2010 #11


    User Avatar
    Science Advisor

    I saw a car today with the license plate GXF132.

    Of all the license plates, what was the odds that I would see that one?
    26*26*26*10*10*10=17,576,000. My oh my, that sure was an unlikely event!
  13. Jul 2, 2010 #12


    User Avatar
    Gold Member

    Were you to pick a plate number at random, or program your computer to do so, and select the GXF132 plate, then set out on a drive and find that plate, it would be more unlikely.

    By your need to be condescending and fatuous, I conclude that you are a liberal progressive Darwinist. Was that a lucky guess, or what?
  14. Jul 2, 2010 #13
    As I understand it homology for chromosomes is defined by phenotypes, not genotypes. So there is a certain region on chromosome A(maternal) coding for blue eyes, and a homologous region on chromosome A(paternal) coding for green eyes. They are two alleles of the same gene, and they are located in the same locus of their respective chromosome.

    Usually, crossovers do not result in "new" proteins only in a recombination of existing proteins. So assume you have again chromosome A(maternal) coding for blue eyes and black hair and chromosome A(paternal) coding for green eyes and blond hair. The whole point of crossover is that there might be sperm cells which contain a chromosome A coding for blue eyes and blond hair.
    I don't know how often this occurs.

    I have to admit that my knowledge on this subject is fairly limited. DNA codes for amino acid sequences and those determine the structure of the proteins. How exactly proteins know how to fold is subject of current research and far from well understood. I suggest you read wikipedia:smile:
  15. Jul 2, 2010 #14


    User Avatar
    Gold Member

    No argument with you on complexity. However, working in several fields has taught me the importance of finding a simplification. That is how physics got its start--- Galileo didn't try to work out how the cosmos operated. He found a problem he could deal with, and ended up developing what became known as Newton's 2nd (have I confused the number?) Law by rolling wooden balls down inclined planes.

    I never enjoy losing a bet, but will enjoy dinner and company, thank you!

    I won't have to check into the Levinthal paradox right yet, because upon reading this, she is wondering if it might contain more munitions for creationism. (Thanks a lot!) I'm sure that she will explain it at the sissy restaurant. If not, count on it that I will check it out. And thanks for tip! Really. I'd never heard of the Levinthal paradox before. I will be digging as deeply as I am capable of digging, and scrounging when necessary. Problem is, I want to figure this out before I die, and I've seen the reaper lurking in the woods.
  16. Jul 2, 2010 #15


    User Avatar
    Gold Member

    Thanks for your replies. On this one, however, I didn't clarify my question properly. I'm aware that protein folding is a mystery, but my question was with respect to the overall structure of an organism.

    So genes make proteins, but what determines where they go? Why are our fingers at the ends of our hands, instead of our feet--- or already pre-inserted into some orifice? Why do our eyes generally end up on our face, instead of the back of our head, which would offer a shorter interconnect? Etc.

    Organisms are clearly more than protein soup. The proteins are carefully arranged. There must be some code which does this. Wikipedia, my favorite source of cross-referenced information, offers nary a clue. Any thoughts about where to go? Does the information I'm looking for exist?
  17. Jul 2, 2010 #16


    User Avatar

    Staff: Mentor

    I understand from your posts that you are in the creationist/Intelligent Design Camp, but you're treading a fine line with violating our guidelines for insults.
  18. Jul 2, 2010 #17
    You might want to read up on the distinction between ontogeny and phylogeny (in case you're not yet familiar with it)

    The development of an individual from a fertilized egg (ontogeny) is a superbly controlled process, also in my opinion far from well understood. The main feature certainly is that in different cells, at different times different genes are expressed (which means translated into proteins) and thus functionally different cells develop. Often there is a complex interplay between different cells, one structure inducing expression of certain genes in another structure. For example the development of the eye in mammalian embryos is induced by the optical nerve growing towards what will be the face. If that optical nerve structure is transplanted somewhere else, the eye will grow there. (In principal at least , I am not a biologist, nor an embryologist).

    So I think it is fair to say that genes encode structure by encoding a certain time-line for ontogeny and a sophisticated interaction between developing structures.

    How that time-line came into being is a different question which is addressed by phylogeny.
  19. Jul 5, 2010 #18


    User Avatar
    Gold Member

    I shall consider myself suitably chastised and be more careful when responding to snide insults in the future. I'll also assume that such insults are strictly the prerogative of posters with titles, although I'd have expected better of a "science adviser." Thank you.

    I am not in the creationist/Intelligent Design camp as you have assumed. That would require believing in the irrational concept of an omnipotent God, which I gave up decades ago. After having been falsely taught as a child, I've chosen not to fill any holes in my understanding of things with a different set of irrational beliefs.

    I've not paid much attention to Darwinism until recently, and these replies to my query have been enlightening. I've heard that it operates more like a belief system than a science, dogmatically intolerant of contrary views, but had dismissed those complaints until my experience here. As a result I have just moved to the "None of the Above Camp," and will devise a better alternative.

    I confess surprise, and considerable dismay, at discovering, as your assumptive statement implies, that this is a politically correct forum.
  20. Jul 6, 2010 #19


    User Avatar
    Homework Helper

    "I am beginning to fear that evolutionary biology is not actually a science. "

    Perhaps you should ask some biologists who do research in this area, and stay away from the highly dishonest creationist camp (by the way, there is no need to write creationism/Intelligent Design, as they are one in the same.
  21. Jul 6, 2010 #20


    User Avatar
    Gold Member

    You are the 2nd person to accuse me of, or imply that I'm a creationist. Kindly work up enough integrity to actually read my posts before labeling me. And why do you need to do that anyhow?

    I started this thread to gain information and perhaps ideas, not to become the target of individuals with personal problems. I notice that while you freely hand out implicit insults, you did not contribute to the question.

    I tried the evolutionary biology literature and could not find estimates for the probability of the evolution of so much as an amoeba. The question of whether or not Darwinism is an adequate broad-scale explanation for biological evolution has been dismissed. So I performed my own basic calculation on the probability of one small human gene coming into existence by random mutation or addition of DNA base pairs, and came up with a probability of 1.4 x 10e-542. That's for one of the 23,000 genes in the human genome, and that's a conservative calculation.

    Such numbers render Darwinist explanations of evolution just as absurd as those of religionists.
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook