Wednesday, August 13, 2014

Flunking the Behe challenge!

Apparently I flunked the Behe challenge [Laurence Moran's Sandwalk Evolves Chloroquine Resistance]. Let's review what happened.

In his book, The Edge of Evolution, Michael Behe calculated the odds of a malaria parasite developing resistance to chlorquine by assuming that two separate mutations were necessary. Here's what he said on page 57 ...
How much more difficult is it for malaria to develop resistance to chloroquine than to some other drugs? We can get a good handle on the answer by reversing the logic and counting up the number of malarial cells needed to find one that is immune to the drug. For instance, in the case of atovaquone, a clinical study showed that about one in a trillion cells had spontaneous resistance. In another experiment, it was shown that a single amino acid change at position number 268 in a single protein, was enough to make P. falciparum resistant to the drug. Se we can deduce that the odds of getting that single mutation are roughly one in a trillion.
There's more, but let's pause here to describe the Behe challenge. Here it is in his own words.
Talk is cheap. Let's see your numbers.

In your recent post on an earlier reviews of my book The Edge of Evolution you toss out a lot of words, but no calculations. You downplay FRS Nicholas White's straightforward estimate that—considering the number of cells per malaria patient (a trillion), times the number of ill people over the years (billions), divided by the number of independent events (fewer than ten)—the development of chloroquine-resistance in malaria is an event of probability about 1 in 1020 malaria-cell replications. Okay, if you don't like that, what's your estimate? Let's see your numbers.
I decided to take the Behe Challenge: Taking the Behe challenge!. The idea was to try and calculate the probability of chloroquine resistance arising in a population of malarial parasites. I don't doubt that this is a rare event but I have serious doubts about Behe's calculations.

On page 66 of his book, Behe claims that the mutation rate is 10-8. He says that the DNA replication machinery makes about one mistake in every hundred million nucleotides. If that's all there is to the calculation then the odds of getting resistance to atovaquone should be 10-8. So where does Behe get one in a trillion (1012)? In my response to the challenge, I pointed out that the overall mutation rate was actually 10-10 or two orders of magnitude lower than what Behe claims. (Did he forget about DNA repair?)

I also pointed out that there's a big difference between the mutation rate and whether a resistant strain becomes established enough to be detected. It's the different between mutation and fixation. Behe seems to be confused about this difference. He seems to think that all you need to do is calculate the probability of the mutations occurring.

Let me emphasize, again, that I'm not questioning the fact that mutations are rare events. What I'm doing is answering Behe's challenge to relate that fact to what we know about mutation and evolution. It turns out that Behe's calculations are suspect. It also turns out that the actual estimates are very much more difficult than Behe imagines.

Let's look at the rest of Behe's paragraph from page 57 in his book.
On the other hand, resistance to chloroquine has appeared fewer than ten times in the whole world in the past half century. Nicholas White of Mahidol University in Thailand points out that if you multiply the number of parasites in a person who is very ill with malaria times the number of people who get malaria per year times the number of years since the introduction of chloroquine, then you can estimate the odds of a parasite developing resistance to chloroquine is roughly one in a hundred billion billion. In shorthand scientific notation, that's one in 1020.
No, that's NOT the odds of a parasite developing chloroquine resistance. It's the odds of the mutations occurring times the odds that it will become well enough established to be detectable.

I tried to meet the Behe challenge by showing how I would calculate the odds of the mutation(s) occurring and why my calculations differ from Behe's. It's still a long shot but that doesn't mean that Behe's calculations and assumptions are correct.

Behe didn't like my answer. He says,
Moran doesn't seem to actually have much confidence in his own numbers. He asks the readers of his blog to help him correct his calculations—which is a commendable attitude but makes one wonder, if he's so unsure of the likelihood of helpful combinations of mutations, whence his trust in mutation/selection? In response to the commenter who alerted him to the huge number of parasites in a million people he writes, "This is why meeting the Behe challenge is so difficult. There are too many variables and too many unknowns. You can't calculate the probability because real evolution is much more complicated than Behe imagines." But, again, if he thinks everything is so darn complicated and incalculable, on what basis does he suppose he's right?
I don't suppose that I'm right for many reasons. For one, I don't know the probability of a chloroquine-resistant parasite surviving transmission through the mosquito host. I don't know whether the various intermediates are more fit than their ancestors of less fit, and by how much. I don't know the frequency of the various alleles in the parasite population. I don't know whether all the humans in an infected region need to be treated constantly with chloroquine in order to detect a chloroquine-resistant strain.

What I do know is that Behe's calculations are wrong. On what basis does he think he's right?


108 comments:

  1. Having observed a phenomenon ten times in half a century, it seems rather silly to declare that it is impossible. Particularly when one doesn't know how many necessary variants are typically floating around in the population, or whether there are enabling or compensating variants that make detrimental mutations less detrimental.

    It just seems silly to calculate odds against something that has already happened, not knowing the things you need to know to do the calculation.

    ReplyDelete
    Replies
    1. Exactly. One can, after the fact, 'calculate' nearly any event into the realm of impossibility (by the common application - incorrect application, as I have been told - of Borel's theorem), as I easily did for a creationist's very existence a few years ago. He dismissed my calculations, but could not explain why Behe's or any similar anti-evolution probability calculations merit.

      Delete
    2. Can you explain in layman's terms how you can use Borel's theorem to make any event improbably?

      Delete
  2. Hi Larry,

    Behe's question, "... if he thinks everything is so darn complicated and incalculable, on what basis does he suppose he's right?" needs to be seen in context. He's not asking why you suppose you are right about your calculations. He's asking why you suppose you are right when you say the major mechanism of evolution is random genetic mutations. As Behe says in the very next paragraph:

    "That's the reason I issued the challenge in the first place. In my experience almost all Darwinists and fellow travelers (Professor Moran doesn't consider himself a Darwinist) simply don't think quantitatively about what their theory asks of nature in the way of probability. When prodded to do so, they quickly encounter numbers that are, to say the least, bleak. They then seem to lose all interest in the problem and wander away. The conclusion that an unbiased observer should draw is that Darwinian claims simply don't stand up to even the most cursory calculations."

    ReplyDelete
    Replies
    1. It's Behe's calculations that don't stand up to serious scrutiny: http://sandwalk.blogspot.com/2014/08/taking-behe-challenge.html?showComment=1406936174596#c9162624727500174727

      Delete
    2. If the calculation Diogenes did was even remotely close to correct, then the statement:
      "When prodded to do so, they quickly encounter numbers that are, to say the least, bleak."
      is about as incorrect as you can get.

      Delete
    3. Hi UE,

      For two neutral mutations, Diogenes' calculations were 1 in 10^14.6. If the first mutation was deleterious, then his calculations were about 1 in 10^18. Microbes can overcome those sorts of numbers, because they have very large population sizes. But larger animals, with much smaller population sizes, couldn't. And if more than two neutral mutations are needed, it begins to become bleak for microbes as well.

      So the question is, why should we be sure that random mutation was in fact the major mechanism of evolution?

      Delete
    4. Hey Bilbo, do the numbers. Don't just make these vague claims:
      "But larger animals, with much smaller population sizes, couldn't".

      Couldn't do what? In what timescale? Do the calculations to back up your claims.

      "And if more than two neutral mutations are needed, it begins to become bleak for microbes as well."

      Do the numbers then, prove these claims. Show us that you are right.

      Delete
    5. And don't forget (as many seem to) the role of recombination in these sexual species. If it requires mutations A and B together, these must be separate mutations at some map distance apart. It is not simply a matter of probability of A arising in the B subpopulation + probability of B arising in A, but also the probability of a recombinational event between these subpopulations. And then add in the fact that there may also be mutations C, D, E, F, G etc which, in combination, may produce resistance. If we know A and B are the only ways to get resistance, then fine. If we don't, we simply can't get a handle on the probability of resistance arising.

      Delete
    6. I did a bit of Bilbo's "bleak for microbes" super-simplified, back-of-the-envelope calculation and attempted to extrapolate the four chloroquine resistance mutations in a decade, over the entire history of life (3.5 billion years).

      If something that requires 4 sequential mutations takes a decade to evolve in a population of 10^12, how much evolution by point mutations only, is possible in 3.5 billion years, in a total population of about 10^30 cells(a rough estimate of the total number of cells on earth)?

      Well, 10^30 is 10^18 times larger than 10^12.

      3.5x10^9 years/10 years = 3.5x10^8 total "decades", each of which evolves 4 mutations.

      How many total mutations have evolved then?
      10^18 * 3.5x10^8 * 4 = 1.4x10^27 total mutations.

      Shit, looks like more than enough to evolve basically every protein and regulation site known to man, and then some. And we haven't even factored in things like gene duplications, frameshifts, polyploidy and so on.

      Wasn't Behe's calculation supposed to make us conclude that the evolution of chloroquine resistance is so slow that we should infer that larger, more complex things couldn't possibly evolve?

      Isn't that what Bilbo was trying to insinuate with his "even looking bleak for microbes" assertion?

      Delete
    7. Bilbo, Behe was lying when he described the results of my calculation. You copied Behe, what Behe said about my calculations, instead of looking at my actual calculations.

      Do not trust Behe or what he says about me! Look at my actual math!

      Behe quotes my figure of 818,500, but he lies about what it measures. I computed that for 50 instances, that's 50, of a double mutant when BOTH single mutations are LETHAL. Behe says my number refers to ONE appearance of a double mutant when ONE of two single mutants are DELETERIOUS.

      DO NOT TRUST INTELLIGENT DESIGN PROPONENTS.

      Delete
    8. But larger animals, with much smaller population sizes, couldn't. And if more than two neutral mutations are needed, it begins to become bleak for microbes as well.

      So the question is, why should we be sure that random mutation was in fact the major mechanism of evolution?


      And then there's the ugly point that Behe never acknowledges, which is that he is assuming that there is one and only one possible protein that can do the job, as if evolution was teleological and had to be pursuing specific unique outcomes. Point is, without knowing how many possible protein outcomes can do a particular job, it's pointless to calculate the probabilities. One can calculate the probabilities of a particular outcome, as in this case, but that does not imply that that outcome is the only one. If that was the case, every time there's a landslide and a stone lands in a particular place, the probability for landing at that particular place in a particular orientation would be so low that one would have to invoke god all the time. It's obviously stupid, and so is Behe's assumption. But he knows that.

      So the answer to "why should we be sure that random mutation was in fact the major mechanism of evolution?" is that it fits the experimental and observational rather well, the theoretical mechanism makes sense, and so far we have seen no major probabilistic "edge of evolution".

      Todd Wood could teach a thing or two to Behe.

      Delete
    9. And then there's the ugly point that Behe never acknowledges, which is that he is assuming that there is one and only one possible protein that can do the job, as if evolution was teleological and had to be pursuing specific unique outcomes.

      I think you're restricting the universe of outcomes too much even here. The question is actually how many proteins would do some job, not just a specific job. Some jobs never have proteins evolve to do them, and those we don't see or think about. Ascertainment bias. We're looking at all the arrows that hit the barn, and then painting bullseyes around them.

      Delete
    10. Oh, the restriction was intentional, since I was just commenting on the particular case at hand. What you say is of course correct. Behe's logic is that if we rewind the history of life and hit the start button again, we would get exactly the same proteins, in exactly the same order (not to mention the big finale, Man). He also ignores that functionality of a particular protein also depends entirely on what other proteins it has to interact with, and so forth. In other words, we have no idea what the sequence space is for every possible scenario.

      Delete
    11. Here is a link to my calculation where I showed Behe was wrong by a factor of ~10,000. This is the calculation that Behe refers to, but does not describe accurately. Read it yourself and compare it to what Behe says I said.

      Delete
    12. Another thing is that Behe ignores that different cells may evolve different resistance mechanism for the same toxic. For example, one cell may develop a protein that inactivates the compound while another develops a system to pump it out of the cell instead. We know a few, but how many possible strategies are there? How on Earth can you make a probabilistic statement about CR overall when you are restricting yourself to one protein solution only? It's the landslide rock scenario all over again. Behe also, as others pointed out, takes a number of around ten observations of known CR as if those cases were all there is, and then uses that number. Does he have a crystal ball? And then he implies that other people are too dumb because you only need a "back of the envelope" calculation to show ToE is "impossible" (read "that god exists"). As Moran would say "and you wonder why we call them IDiots".

      Delete
    13. Bilbo,

      1 in 10^20 is the number of cell replications required for chloroquine resistance to reach detectable levels. Behe is wrong in saying that that's the number of replications needed for two mutations to occur in one cell. If the mutations occur in only one cell, it won't be detectable. The mutations have to spread through the population and get fixed before chloroquine resistance is detected. Therefore, the required mutations will occur a lot faster than Behe imagines.

      Delete
    14. Bilbo,

      Michael Behe has spun and cherry picked Diogenes' numbers.

      Diogenes said:

      If mutations can occur "sequentially" in one organism, and s = 0.01 for the double mutant, 21,000 people must become infected before the double mutant gets fixed and established in the population.

      If mutations must occur "simultaneously" in one organism, and s = 0.01 for the double mutant, 818,500 people must become infected before the double mutant gets fixed and established in the population.

      Now, Behe ignored the first calculation and cherry picked the second calculation because it's a higher number, although it's clear that the mutations don't have to occur simultaneously.
      Furthermore, those values show the number of people that have to be infected for the mutations to get fixed and established, not simply for the mutations to occur in one cell.

      Delete
    15. Thank you Vimal. The numbers quoted above assume each type of mutation occurs FIFTY (50) times in the population. Behe presented it as each mutant occurring ONE TIME, and then compared that against his shit figure of 10^20 which refers to ONE instance of a double mutant. He compared my apples against his oranges. Sneaky sneaky.

      Moreover, the 818,500 refers to a double mutant where BOTH single mutants would be LETHAL. Behe tells his audience I computed that for the case where ONE single mutant would be deleterious. Dead wrong. For that case, I actually did do a calculation, but I didn't post it--- the actual number is about 1 in 1,000 for one instance of a double mutant, or 1 in 50,000 for 50 double mutants. Behe' s figure of 10^20 means one double mutant happens in 10^8 patients, I showed it is 10^3, so even under this assumption, Behe is wrong by a factor of 10^5.

      Compare apples to apples. Behe tries to conceal his blunders by comparing my apples to his oranges-- and then lying about how many apples I have!

      All Behe's many blunders point in the same direction. That nonrandom pattern can't be produced by blind chance, it must be his deliberate, purposive, intelligent design.

      Delete
    16. Thou shalt not bear false witness against thy neighbor

      Behe has been forgetting about this one for the past 20 years.

      Delete
    17. From here:

      13
      Larry MoranAugust 14, 2014 at 3:22 pm
      We all agree that some genotypes, like chloroquine resistance in Plasmodium, are extremely rare. That’s not an issue. We also agree that such combinations of mutation will be beyond the reach of some species.

      Not being much of a mathematician, I would just hazard a guess that species that have populations significantly below 10^14.6 would find the chances rather bleak.

      Delete
    18. Bilbo,

      You cannot apply one case to the whole of evolution. Yes, there will be easy mutations, hard mutational combinations, more likely scenarios, less likely scenarios etc etc. This is exactly what we would expect in a natural world. If there was a divine hand in the proceedings, all events, easy or hard, should have been equally probable.

      Delete
    19. VR: You cannot apply one case to the whole of evolution.

      Moran: We also agree that such combinations of mutation will be beyond the reach of some species.

      Moran doesn't say which species, but am I unreasonable in thinking species whose populations are too small?

      VR: If there was a divine hand in the proceedings, all events, easy or hard, should have been equally probable.

      Why?

      Delete
    20. Moran doesn't say which species, but am I unreasonable in thinking species whose populations are too small?

      No, that is what Moran means, unless I missed something.

      Delete
    21. This comment has been removed by the author.

      Delete
    22. Hi Pedro,

      I misread your comment the first time through. Okay, so my understanding of the issues isn't completely way off.

      Delete
  3. Hi Petrushka,

    Behe isn't saying the evolution of chloroquine resistance was impossible.

    ReplyDelete
  4. Behe says chloroquine resistance did evolve. His most important claim is that CR is a reasonable model for many or most of the protein-protein interactions within a cell. Because it required on the order of 10^20 cells for this one interaction to evolve this means that most of the p-p interactions in a eukaryotic cell ( or animal or vertebrate) could not have evolved

    ReplyDelete
    Replies
    1. Agreed. IIRC Behe uses his CR numbers to claim that no multi-protein complex could have evolved in the mammal ( or primate?) lineage based on estimates of the total number of individuals in that lineage. He neglects to give an example of such a complex that he claims couldn't have evolved ( I cant think of any)
      I wish someone would critique his claims about protein protein interactions. He presents a picture similar to the cartoons you'd see in a textbook, of protein complexes and networks all interacting specifically with the implication that a new interaction would be astronomically unlikely. I think many of the interactions in signaling networks involved low-complexity sequences and are hence easier to evolve. More to the point, I recently saw a reference which reported that a significant fraction ( 30%?) of p-p interactions are non-functional. A paper I recently skimmed ( Sarah Taichmann lab, Cell) showed that the surfaces of proteins evolve to minimize stickiness and that stickiness negatively correlates with protein abundance. If novel p-p interactions were so unlikely its hard to see how this would be the case.
      Forgetting ID for the moment, it seems to me this phenomena is in the same category as non-functional TF binding sites, ectopic gene expression, non-functional splice forms ( which I heard about here a few weeks ago) and moonlighting enzymes ie. examples of how living systems randomly explore neutral 'complexity space'

      Delete
    2. There was a lot on this back in 2007. E.g. my TREE review, or: https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=musgrave+%22An+Open+Letter+to+Dr.+Michael+Behe%22

      Delete
    3. And: http://www.pandasthumb.org/archives/2008/04/behe-versus-rib.html

      Delete
  5. "It's the odds of the mutations occurring times the odds that it will become well enough established to be detectable."
    Or, to bind back to the phrasing "appeared less than ten times" which should have been "observed to appear less than ten times", it's the odds of the mutations occurring times the odds that it will become well enough established to be detectable times the odds that the detectable mutation is actually detected.

    Not everyone who contracts malaria gets to try chloroquine and not everyone who does gets to participate in a follow-up study.

    ReplyDelete
    Replies
    1. And published. Behe assumes every parasite that.ever had a double mutant established a subpopulation, got somebody sick who was already on chloroquine, got detected, AND was published in the literature. How many bush doctors have the time?

      Delete
    2. Behe doesn't asume anything. He's doing missionary work, and I can hear the acolytes cheering "victory!" at UD.

      Delete
    3. Pedro, of course they are:
      Evolutionary biologist Larry Moran tries calculating with big numbers re evolution.

      I would personally recommend this post for the Outstanding Ignorance and Stupidity Award, hard as it is to nominate the winner.

      Delete
    4. I'd bet on that horse Piotr. They don't realize it, but with their (completely speculative) numbers they are actually computing the probability that a horse could arise in an act of sudden creation - and the numbers dont look promising to say the least. I'd say they are casting doubt upon the god hypothesis but that would be the scientific view which is not something to be found in Awake! magazine.

      Delete
  6. So what is he really proposing? That the Infant Jesus personally intervened ten times in the last half century to make malaria harder to treat? And he wants to worship the thing that did that?

    ReplyDelete
  7. Off topic (except as to IDiocy), has anyone seen this over at Uncommon Descent, by "johnnyb," entitled "Do Darwinists Think that Women are Closer to Chimps than Men?" Neil Rickert has a report here:

    https://nwrickert.wordpress.com/2014/08/07/id-thinking/

    It is Ray Comfort-like in that it all but commits the fallacy that human males and females had to evolve separately. He even mentions bananas.

    ReplyDelete
    Replies
    1. "it all but commits the fallacy that human males and females had to evolve separately. He even mentions bananas."

      To paraphrase Mae West: Are those bananas, or are the human males just happy to see the human females?

      Delete
  8. Larry:

    you write:

    No, that's NOT the odds of a parasite developing chloroquine resistance. It's the odds of the mutations occurring times the odds that it will become well enough established to be detectable.

    If a trillion replications take place in any one person, with Chloroquine causing the disease to fall off, would you agree that, at the very least, the likelihood of CR is no less than one in a trillion?

    What I do know is that Behe's calculations are wrong. On what basis does he think he's right?

    What calculations? These are nice sounding words to someone who is ready to disbelieve Behe and ID arguments, but you seem to miss the point that in all of this Behe makes NO calculations.

    First, he quotes White's estimation. Should you want to disagree with these numbers, that's fine. But don't offer the canard that represents "Behe's calculations."

    Again, when he says that atovaquone resistance occurs once every million replications, this is NOT a calculation. It is an observation. Here's what Behe wrote (from above):

    For instance, in the case of atovaquone, a clinical study showed that about one in a trillion cells had spontaneous resistance.

    You're not disagreeing with "Behe's calculations"; you're disagreeing with nature.

    ReplyDelete
  9. Larry:

    A follow-up:

    You write:

    I don't suppose that I'm right for many reasons. For one, I don't know the probability of a chloroquine-resistant parasite surviving transmission through the mosquito host. I don't know whether the various intermediates are more fit than their ancestors of less fit, and by how much. I don't know the frequency of the various alleles in the parasite population. I don't know whether all the humans in an infected region need to be treated constantly with chloroquine in order to detect a chloroquine-resistant strain.

    From a website on malaria:

    Growth and division of each oocyst produces thousands of active haploid forms called sporozoites. After the sporogonic phase of 8–15 days, the oocyst bursts and releases sporozoites into the body cavity of the mosquito, from where they travel to and invade the mosquito salivary glands. When the mosquito thus loaded with sporozoites takes another blood meal, the sporozoites get injected from its salivary glands into the human bloodstream, causing malaria infection in the human host. It has been found that the infected mosquito and the parasite mutually benefit each other and thereby promote transmission of the infection. The Plasmodium-infected mosquitoes have a better survival and show an increased rate of blood-feeding, particularly from an infected host.[3-5]

    source: http://www.malariasite.com/malaria/LifeCycle.htm

    So, if the "oocyst" produces "thousands" of copies of the 'sporozoites,' then the minimum probability of CR moves from 1 in 10^12 to 1 in 10^15.

    Things look bad. Notice that 'fitness' improves. Notice, from the website, that there is NO indication of any problem with the 'sporozoite' "passing through" the mosquito.

    Which only leaves the question of "alleles" and widespread usage of Chloroquine.

    As to "alleles," it would seem that since we're dealing with asexual reproduction going from one person to the next, it matters little what the population as a whole looks like. Or so it seems to me. Maybe you could give some reasons for looking at it otherwise.

    As to widespread usage of Chloroquine, let's say that only one out of ten people use Chloroquinone. But what if ONE mosquito bites ten people? Then everything averages out, doesn't it?

    When knowing something helps the cause, then Darwinists are very certain; and when not knowing something is helpful, then they tell you how hard it is to know something.

    Unsavory.

    ReplyDelete
    Replies
    1. Lino, if not via evolution, then what is the cause of malaria, and why?

      Did/do yhwh-jesus-holy-ghost create and inflict unsavory, deadly diseases and parasites upon people and animals because the first specially created dust-man and rib-woman ate some forbidden fruit?

      Delete
    2. As to "alleles," it would seem that since we're dealing with asexual reproduction going from one person to the next, it matters little what the population as a whole looks like. Or so it seems to me. Maybe you could give some reasons for looking at it otherwise.

      Well, the population as a whole becomes enriched in a particular type, even in fully asexual species, due to drift/selection. If that type happens to be a potentiating mutation, it increases the odds. But these aren't asexual anyway. "In the mosquito's midgut, the gametocytes develop into gametes and fertilize each other,"

      Delete
    3. The vast majority of replications are asexual, in the ratio of around a million to one.

      Delete
    4. Lino: "So, if the "oocyst" produces "thousands" of copies of the 'sporozoites,' then the minimum probability of CR moves from 1 in 10^12 to 1 in 10^15. "

      No, you are multiplying when you should actually add. It becomes 10^12 + 10^3 =~ 10^12. You stupidly computed 10^12 * 10^3. You are wrong by a factor of 10^3.

      Deal with my numbers. Behe lied about the results of my calculation, which was very clear.

      If both single mutants are neutral, a double mutant should appear in every few hundred people after only ONE round of infection. Say 10^3 (being generous) * 10^12 = 10^ 15. So Behe is off by 10^5.

      Here is my calculation Behe lied about. If BOTH single mutants are LETHAL, a double mutant appears in one out of 20,000 people after only ONE round of infection. 10^4 * 10^12 = 10^16 and Behe is wrong by 10^4.

      He did not tell the truth about my calculations.

      Delete
    5. Lino: a malaria infection involves 30 or 40 rounds of replication. At least one replication in the mosquito is sexual. Your ratio of "a million to one" is wrong by a factor 10^3 to 10^4.

      For the calculating how common a rare mutant is, what matters is how many rounds of replication there are, not how many cells are immediately descended from sexual reproduction. In several rounds of infection, sexual reproduction might occur about every month, ballpark.

      Delete
    6. Lino - the vast majority of replications in the human germ line (even reining in your 'millions' hyperbole for Plasmodium, corrected above) are asexual as well. The intermittent nature of syngamy/recombination does not prevent it from making a 'population' in the gene pool sense.

      Delete
    7. Diogenes:

      I didn't say that the ratio of "rounds of replications" between asexual and sexual are a million to one, I said that the number of "replications" is a million to one.

      If a trillion parasites develop in the average human infection, then, since reproduction is asexual and direct, then half as many individual copies of the parasite are needed. So 0.5 x 10^12 replications take place in the last round. I should have said the ratio was a billion-to-one. since to get a thousand sporozoites in the oocyst, the generation before had to number 500.

      Delete
    8. Alan Miller:

      Malarial parasites have separate existences. The genome from one human doesn't mix with the genome of another human. Although there is sexual reproduction within the mosquito, it doesn't appear that the genome of one line is mixed with that of another.

      This isn't the case with humans. Yes, meiosis occurs. But then these copies combine, one coming from the male and from the female. And it's possible for 'alleles' to move around in the population. But with mosquitoes, each lineage, as far as I can see, is entirely separate from every other line. And so the point I was making was that what goes on in other mosquitoes has no relevance to what happens in the mosquito that is carrying a parasite with CR, or not.

      Delete
    9. Lino:"So, if the "oocyst" produces "thousands" of copies of the 'sporozoites,' then the minimum probability of CR moves from 1 in 10^12 to 1 in 10^15."

      Where does the 1 in 10^12 come from?

      Delete
    10. "But with mosquitoes, each lineage, as far as I can see, is entirely separate from every other line. And so the point I was making was that what goes on in other mosquitoes has no relevance to what happens in the mosquito that is carrying a parasite with CR, or not"

      Not according to population genetic studies of malaria. There is extensive lineage mixing of malaria parasites across much of their range.

      Delete
    11. Sorry Lino, you're wrong. Different Plasmodium lineages do meet each other in a single mosquito, and recombine.
      Just one example: Recombination in Plasmodium

      Delete
    12. Chris B:

      If the parasite must pass from one person to another before CR develops, then, owing to the trillion parasites that develop in the human host, the probability of CR developing is less than one in a trillion = 1 in 10^12.

      Delete
    13. Alan:

      The study you point to concludes its abstract with this sentence:
      "This study supports the feasibility of genome-wide association studies in some parasite populations."

      Put another way, "genome-wide association studies in parasite populations . . . is not feasible."

      My background in this is slight and shaky, but they seem to be saying that LD--linkage disequilibrium--cannot be established between different lines because there aren't enough "markers." But, using indications from "recombination," the possibility of zeroing on useful "markers" is "feasible."

      This doesn't dispute the point I've made. Just think it through some.

      And just think of this: in one person infected with P. falciparum, 10^13 new organisms develop. That is more than the total number of mammals that ever existed. You have to admit that each infection represents some kind of 'lineage'!

      Delete
    14. Diogenes:

      About your "calculations."

      (1) According to your calculations, CR should happen almost immediately. 420 people is nothing compared to the millions that are infected each and every year. So, something's wrong with your numbers.

      (2) When you made your calculation, you decided not to include the effect of "back mutations." This does affect 'actual' probabilities.

      (3) Per your calculations, even though we would expect to find ONE "double-mutant" given, on average, 420 individuals, the odds of that parasite entering into the mosquito would be miniscule, roughly, I would surmise, the ratio of the volume of blood the mosquito ingests, over the total volume of blood present in the infected individual. Coupled with (2), this certainly affects the probabilities, etc.

      (4) Per your numbers, the probability of getting a "double mutant" is 10^12 divided by 420 = 2.5 x 10^-15. So, per neo-Darwinian theory, as given by your calculations, to arrive at a 2 a.a. change where BOTH mutations are simultaneously present, requires almost a million billion replications.

      Is that the best that neo-Darwinism can do?

      And isn't this Behe's point exactly?

      P.S. I won't mention again that Behe "calculated" nothing. It took P. falciparum, per White's numbers, 10^20 replications to confer CR. What the great surprise is not so much the big numbers, but the LITTLE numbers: i.e., that with all of this evolutionary firepower made available, when we investigate the genome of the P. falciparum all we find is a 2 a.a. change. Not a 6 a.a. change. Not a 10 a.a. change. Not a 20 a.a. change. That's the surprise. And it should make it easily clear that neo-Darwinism cannot be invoked to explain evolution on any scale other than the very tiny.

      Delete
    15. Lino,

      ///I won't mention again that Behe "calculated" nothing. It took P. falciparum, per White's numbers, 10^20 replications to confer CR. What the great surprise is not so much the big numbers, but the LITTLE numbers: i.e., that with all of this evolutionary firepower made available, when we investigate the genome of the P. falciparum all we find is a 2 a.a. change///

      The likelihood of CR is not the likelihood of 2 mutations occurring. Behe and you are wrong.
      For resistance to reach detectable levels, mere appearance of the mutations is not enough. The required mutations must originate and then spread through the parasite population and reach a high frequency.

      Read the new Summers et. al., paper.

      http://www.pnas.org/content/111/17/E1759.long

      2 mutations only confers low resistance. Even at moderate doses of chloroquine, cells with 2 mutations may not survive. 4 mutations are needed for significant resistance. Moreover, some combination of these mutations decrease fitness in the absence of chloroquine. So mutant cells will be outcompeted by their wild-type relatives when the drug is absent. This will affect the spread of mutant parasites within a population since exposure to chloroquine will fall when the parasites are taken up by mosquitoes. There are other pitfalls too. For instance, if K76 gets mutated to any amino acid other than T, the ability to transport chloroquine weakens even if all the other required mutations are present.

      Given all these factors, CR is difficult to develop if treated with a sufficient dose of chloroquine. The parasite population will be eradicated before 4 of the right mutations occur and get a chance to spread at high frequency.

      Extrapolating this to reach a blanket verdict against all of evolution is a total fallacy, since the likelihood of mutations occurring and spreading are different in different cases and contingent on many factors.

      Delete
    16. Lino says:"If the parasite must pass from one person to another before CR develops, then, owing to the trillion parasites that develop in the human host, the probability of CR developing is less than one in a trillion = 1 in 10^12"

      That makes no sense. Why does the parasite have to pass from one person to another before CR develops? Are you saying that if a population of a trillion parasites grows in an individual, that at most one of these will develop CR?

      In any case, the probabilities you are discussing undermine Behe's calculations; they do not support them. Behe's major mistake is that he assumes the ten or so instances of CR that have been documented represents all instances of the evolution of CR in P. falciparum. The chance of developing CR in Behe's calculations, 1 in 10^20, is 10 (or so) divided by the total number of malaria parasites that have grown in all human victims over the study period. This makes an assumption: That those 10 instances of documented CR are the only instances ever to have evolved in the history of the species. This is a completely unsubstantiated assumption. It also necessitates at least 3 corollary assumptions:

      1. That all of those cases of CR arose in a population exposed a level of treatment with cloroquine such that CR rose to sufficient levels in the parasite population to be detected by health workers and documented to the point it could be published in peer-reviewed literature. Otherwise those instances of CR could not have become part of Behe's numerator in the 10^-20 calculation.
      2. No instances of CR arose in a mosquito, or else
      3. Every instance of CR that developed in a mosquito went on to infect a human who then went on to fulfill the requirements of 1.

      Things are looking bleak for Behe's hypothesis.

      Delete
    17. Lino says:"Per your numbers, the probability of getting a "double mutant" is 10^12 divided by 420 = 2.5 x 10^-15. So, per neo-Darwinian theory, as given by your calculations, to arrive at a 2 a.a. change where BOTH mutations are simultaneously present, requires almost a million billion replications"

      No. Not that both mutations are simultaneously present. That probability represents the probability that both mutation occured simultaneously in the same individual parasite on the same DNA molecule. We already know empirically that this is not required for that partucular genotype of CR to appear.

      Delete
    18. "And just think of this: in one person infected with P. falciparum, 10^13 new organisms develop. That is more than the total number of mammals that ever existed."
      Nonsense. Probably the combined number of rats and mice alone (and related murids), that have ever existed, is substantially bigger than ten trillion.

      Delete
    19. This doesn't dispute the point I've made. Just think it through some.

      I'd be grateful if you didn't patronise me! ;) The point you made was that reproduction in P falciparum was asexual. You wheeled that back to 'it's asexual a million times more often than sexual'. That was corrected also. So now, you think recombination is insignificant because genome-wide association studies are not currently being done for resistance genes, and there's more LD than you'd expect if recombination was maximally efficient and there was minimal epistasis.

      I dunno why this is such a big deal for you - you accuse 'Darwinists' of hiding from inconventient truths in a manner you describe as 'unsavory'. Right back atcha. Why would it be such a big lump to swallow to accept that recombination is a significant factor in P falciparum population dynamics? Does Behe have to be right?

      "We show high variation in recombination rates [...]" bullet point: recombination detectably occurs - we wouldn't be able to say what its rate varied between if it didn't.

      "In natural P. falciparum populations, recombination events will be observed only when a host is infected with multiple genotypes. As a result, “effective” population recombination rates are directly correlated with frequencies of multiple infections, endemicity, and the out-crossing rate"

      Bullet point: recombination demonstrably occurs - though clearly there are factors that reduce its effectiveness, exactly as inbreeding or strong directional selection do in other populations.

      Bottom line: if it occurs, with some frequency, it is a potential source of the AB genotype additional to the B-in-an-A and A-in-a-B serial mutation scenarios. You could model this, if you so chose, and turn the recombination/gene flow 'knobs' to your heart's content. I think you'd be surprised how little outbreeding recombination it takes to make a difference to the occurrence of AB, should both appear in the same geographical area.

      Delete
    20. Vimal R:

      For resistance to reach detectable levels, mere appearance of the mutations is not enough. The required mutations must originate and then spread through the parasite population and reach a high frequency.

      It seems to me like you're talking about 'fixation,' which is a separate issue of CR. If someone infected with malaria is given chloroquine and doesn't recover, then CR has developed. Taking the number of replications in any one human host, and then the numbers of people infected by the parasite before an instance of CR gives you some numbers. White's number was 1 in 10^20. It's not a number Behe calculated. It's nature telling us what it can do and what it can't. Then we, as humans, can interpret those numbers. Behe did that in The Edge of Evolution.

      Delete
    21. Allan Miller:

      I don't see what recombination gives you vis-a-vis the problem at hand. That's one type of 'mutation' if you will, How does recombination change the results of Summers' paper and the numbers we have from White?

      Delete
    22. Lino: "It seems to me like you're talking about 'fixation,' which is a separate issue of CR"

      No, he's not talking about fixation, he's talking about "for resistance to reach detectable levels", which is the only way an instance of CR can make it into White/Behe's numerator.

      Delete
    23. Chris:

      As soon as someone develops resistance, the scientific community is aware of it for fear that such a resistance now exists. Yes, there's some sampling issues involved, but I don't think they change the probabilities by much. When people get malaria, they go to hospitals or receive chloroquine, or both.

      Delete
    24. Lino I don't see what recombination gives you vis-a-vis the problem at hand. That's one type of 'mutation' if you will, How does recombination change the results of Summers' paper and the numbers we have from White?

      Hmmm...

      Sorry if this is a bit tl;dr. But consider a jar containing 10^12 pairs of rods. They have distinct ends, call them L and R. Suppose that L and R start all white. Suppose (unrealistically) that the only way to get CR in this model is a Yellow L and a Black R in the same pairing (not necessarily on the same rod). Suppose that you draw rods at random and either copy one, or discard. And randomly (say once every 10^12 replications) you get a (neutral) BR or YL mutation.

      Suppose matters are such that drift has permitted both BR and YL to increase, independently. Say they each have frequencies of 001% in the population. You can plug in your own numbers if that’s too generous, or give them different proportions; the point is illustrative, not definitive. Take this as a baseline to compare pure asexuality vs asexuality-plus-occasional-recombination. Suppose there are currently no genomes with CR (both BR and YL in the same genome).

      Asexual.

      The chance of a BR/LY combination is the chance of picking a mutant genome and generating the complementary mutation – that’s .002% * 1/((10^12)/2). Giving a remote 1 in 4*10^17, on these numbers.

      Mixed mode – syngamy

      Now we add gametes. The odds for a single pick are the odds of picking a single rod from the population of diploid pairs which has one mutation, and then another which has the other. Mutation rate doesn’t come into this; it is wholly dependent on population frequency of the mutants that have already happened and have left descendants. So, your first pick has a 0.001% chance of being one of the mutations. Your second pick has a 0.0005% chance of being the other (you are picking random haploid chromosomes here). I get that as 1 in 5 * 10^11, 1 in 20 billion syngamies, but 1.25 million times more likely than serial replication alone (which, incidentally, still happens). The smaller the subpopulations, the closer the differential gets to parity. But even with just 200 mutant cells out of 10^12, sex more than doubles the chances of getting the mutations together vs mutation alone.

      Obviously, I have only got the mutants into the same cell, not onto the same chromosome. Sex will just as easily sunder them at the next segregation. But there are many intervening asexual generations, taking place in a milieu which, if it includes chloroquine, will give these genome frequencies a significant boost from selection. It is no longer drift alone that causes increase in frequency. And each increase exponentially increases the likelihood of a successful combination. It’s a bit like the so-called ‘birthday paradox’ – not so much a paradox as a counterintuitive consequence of extra chances. The probability of a pairing does not increase linearly with each extra candidate, but exponentially.

      Mixed mode – crossover

      So now we have a population of CR elements that can increase by selection, when they occur together. They only work in tandem, but they will get together more and more frequently. Every selective gain is a boost for the frequency of both BR and YL. And then, it is simply a matter of waiting for a crossover to weld the elements together, and avoid the losses from segregation. The world is its lobster.

      I would not even know where to begin calculating the above 'properly', since geographic factors and crossover rates should also be included. But calculating a probability is not my goal, just to indicate that you can’t sweep recombination under the carpet should you wish to so calculate.

      Delete
    25. Lilo:
      "When people get malaria, they go to hospitals or receive chloroquine, or both."

      Would that it were so. There are an estimated 250 million NEW cases of malaria each and every year in the world ( a rough estimate). When I was a graduate student 20 years ago, an estimated minimum of 500 million NEW cases of malaria occurred each and every year. Many of these cases happened in remote areas with absolutely no public health system, no available treatment except for traditional remedies, and zero information collected. Many more cases happened with minimal public health intervention, no or minimal drugs available, and minimal scientifically useful data acquired. We go through several more layers of incremental health care/drug/surveillance improvement before we get to people getting hospital and/or drug treatment on a widespread level. I'm sorry if I sound testy at this point, but as someone who is somewhat familiar with the sheer immensity of malaria as an impediment to human well being, I cannot let these assumptions go uncontested.

      To be clear: the chances that the <10 instances of chloroquine resisitance documented in the scientific literature is an accurate representation of the total number of times CR has ever happened in the time period considered by White/Behe is absurd, and completely indefensible, for the reasons I identified above. In other words, the White/Behe numerator has no credible evidence to support it.

      This is (at least in part) what Dr. Moran was trying to get at in his post, that there are so many unknown variables in how CR might arise and be detected that it is pretty much impossible to accurately determine White/Behe's numerator.

      Instead of coming to the unsavory, uncharitable conclusion that Dr. Moran was playing dumb about what we know about CR spread in malaria populations, you might want to think about the myriad ways in which CR appearance could go undetected. Considering the time Behe spent dwelling on how scientists fail to pay enough attention to important details of biochemical complexity in 'Darwin's Black Box', he sure glosses over enormously complex machinations of malaria epidemiology.

      Delete
    26. Allan:

      Thanks for the illustration. However, I think your explanation is simply along the lines I suspected, and that is that the likelihood of arriving at the needed 'combination' of mutations is, if anything, increased by recombination. But if the probabilities are increased---and let's assume slightly---then the numbers we're dealing with only become stronger in pointing out the limited space of possibilities that neo-Darwinian mechanisms can explore.

      That's what I see you saying here:But there are many intervening asexual generations, taking place in a milieu which, if it includes chloroquine, will give these genome frequencies a significant boost from selection. It is no longer drift alone that causes increase in frequency. And each increase exponentially increases the likelihood of a successful combination.

      And that's what I had in mind when I said, "How does recombination change the results of Summers' paper and the numbers we have from White?" I should have probably said, "How does this appreciably change the results of Summer's paper and the numbers we have from White?"

      Delete
    27. I don't understand your dismissal. Yes, of course recombination increases the chances. My whole point. I gave a numerical example, showing that recombination/syngamy only has to happen infrequently to have a substantial effect on the probabilities - in the example, it was a million times more effective than serial mutation at generating a particular double mutant from the same starting population. You can pick other figures to give lower differentials, but recombination nearly always wins. It is much more powerful, therefore you can't just ignore it when calculating the probability of a particular combination.

      But if the probabilities are increased---and let's assume slightly---then the numbers we're dealing with only become stronger in pointing out the limited space of possibilities that neo-Darwinian mechanisms can explore.

      How does its contribution to raising the likelihood constitute an even greater limit on 'neo-Darwinian' mechanisms? ad absurdum, if the likelihood was zero, there would be no limit on the capacity of 'neo-Darwinian' mechanisms! Nah. Syngamy and Recombination make evolution happen faster and allow much more broad probing of the space of possibilities than serial mutation alone. That's why Life went a bit wild after 2 billion years of asexual stasis.

      Delete
    28. It seems to me what you're saying is that the malarial parasite, because of its short period of existence within the mosquito where it can there reproduce sexually, now has an even greater chance to explore the possibilities available to it in finding a way to resist the deadly effects of chloroquine.

      But the entire discussion here is one where I assume that the Darwinists want to "limit" the possibilities of the parasite so as to better explain the very limited extent of its final adaptation.

      What am I missing here?

      Delete
    29. Mikkel Rasmussen:

      Yes, you're probably right. I made a very quick calculation in my head---but the calculation is meant for effect, not to form the basis of some study of mammalian lineages.

      Why don't we take chimps. Six million years ago, it is thought that the split between chimps and humans took place.

      Well, let's say that population of chimps world-wide, at the time, was one million. Let's say that population size was maintained for the intervening six million years. That's six trillion replications. Per what we see happening in the malarial parasite, any change to the chimp genome that required not 'one' step, but 'two' steps (mutations" occuring in one organism in the population at the same time, then not even ONE such change has occurred.

      These kinds of numbers make neo-Darwinism look wholly inadequate.

      Part of the discussion of Behe's work should include the mathematical model that he and Snoke developed. The model predicted that huge numbers of generations would be needed for even minor changes to the genome.

      They laughed at his work. But, of course, laughing is not refuting.

      Nevertheless, Behe then looked for real-life confirmation of the numbers his model had generated. He looked to the malarial parasite because it was so well-known and studied, and, the selective advantage was as favorable as NS could give you.

      And the numbers show that only mild changes occur. Almost next to nothing. And, of course, think of smaller populations of animals throughout the world. How did these possibly evolve?

      He was laughed at again. Told he didn't know what he was talking about.

      And now Summers' paper demonstrates that CR is NOT POSSIBLE without AT LEAST TWO SPECIFIC MUTATIONS! The purpose of the 'caps' is to help you to 'focus like a laser' on the true issue at hand. This paper CONFIRMS Behe's surmise that the malarial parasite, because of the difficulty it has in developing CR is required to "find" TWO mutations, not just ONE. This has now been demonstrated.

      Behe was right. The Darwinists were wrong. Do we hear any apologies?

      No.

      Do we see people changing their views? No.
      Do we see people changing their minds? NO.

      The LHC smashed protons together at such high energies that they could explore particles of very high energy=very high mass. String Theory posits particles of very high mass, so high, in fact, that we hadn't discovered them before the LHC got up to its high energies. But now, even at these high energies, or, equivalently, these high masses, no new particles have been detected.

      Now, the scientific approach would be to say that theory has been severely undermined. Instead---in the field of physics, not biology---in a field that lends itself to exactitude, excuses are now being offered as to why these particles were not detected.

      Eventually, the string theorists will have to give up. And so, too, the Darwinists.

      Behe, per the Darwinian imagination, should not have come up with the numbers and findings that he did. Instead of changing your views, you heap scorn on the "messenger."

      Alas.

      No one has said anything here that takes away from the import of

      Delete
    30. I mangled the last sentence in the third paragraph. It should read:

      Per what we see happening in the malarial parasite, any change to the chimp genome that required not 'one' step, but 'two' steps (i.e., two mutations" occurring in one organism in the population at the same time) then not even ONE such change has occurred.

      BTW, if you say that "no" 'two-step' changes need take place, then, given enough time, there is NOTHING preventing a chimp from becoming a gorilla, or a human to start becoming a chimp.

      I don't know about you, but I don't see anything like that happening.

      Delete
    31. Lino says: And now Summers' paper demonstrates that CR is NOT POSSIBLE without AT LEAST TWO SPECIFIC MUTATIONS! ...This paper CONFIRMS Behe's surmise that the malarial parasite, because of the difficulty it has in developing CR is required to "find" TWO mutations, not just ONE. This has now been demonstrated.

      Behe was right. The Darwinists were wrong. Do we hear any apologies?"


      Behe should apologize for being dead wrong about this, and everything else he's ever said. No scientist said CR resistance required one mutation! Absurd. The scientists arguing with Behe said it might require two or three or four-- the actual number is four-- and if single mutants were neutral, or slightly deleterious, they could be sequential. Behe said it was two and they had to be simultaneous which means each single mutant had to be lethal.

      He said simultaneous. The mutations were sequential. Behe was wrong, again. Behe also said no protein binding sites had ever evolved nor could evolve because functional sites require multiple simultaneous mutations; and he was immediately smacked down by ERV who pointed out that Vpu evolved a protein-protein binding site, plus it's a gated cation channel and it evolved a Golgi-binding targeting sequence. Scientists piled up example after example of newly evolved protein binding sites. Then there was bovine seminal ribonuclease evolving a protein binding site by changing a single amino acid, and the examples compiled by Grueninger et al. There was the alanine scanning mutagenesis work of James Wells in the 1990's, and affinity maturation of antibodies in the immune system.

      Molecular biologists know how easily functional sites evolve, and they have no reaction to Behe's work except to feel douche chills as he face-plants himself over and over.

      Delete
    32. Lino says:
      "But the entire discussion here is one where I assume that the Darwinists want to "limit" the possibilities of the parasite so as to better explain the very limited extent of its final adaptation.

      What am I missing here?"

      You're missing a lot, aaparently. No one is trying to 'limit the possibilities of the parasite'. Quite the contrary, as I mentioned above, many of the limitations you are talking about undermine Behe's calculation. The 'very limited extent' of malaria's adaptation is an assumption og Behe that has zero evidence to support it.

      We have been trying to explain it to you in so many different ways here, that you should not be asking us what you are missing. You should be asking yourself.

      Delete
    33. Hilariously, Lino invokes the "Why don't chimps give birth to humans?" argument, one of the dumbest in all the arsenal of creationism:

      "if you say that "no" 'two-step' changes need take place, then, given enough time, there is NOTHING preventing a chimp from becoming a gorilla, or a human to start becoming a chimp."

      This bears no resemblance to evolutionary theory. To which I reply: if your God is real, there's nothing to prevent him from turning humans into chimps.

      Delete
    34. Lino

      But the entire discussion here is one where I assume that the Darwinists want to "limit" the possibilities of the parasite so as to better explain the very limited extent of its final adaptation.

      The 'Darwinists' should, I hope, wish to be scrupulous about including ALL the factors involved.

      If recombination makes chloroquine resistance more likely, and yet it's still rare, this doesn't invalidate the other points made - the stochastic nature of occurrence, the inevitable excess of occurrence over detection, the possibility of unknown selective effects on the single mutants. The OP was about the many difficulties in making the calculation. But the discussion also brings out the difficulty of extrapolating from a single case to the whole of evolution. It certainly makes no sense to extrapolate from a model that ignores recombination to infer limits on sexual species.

      My ulterior reason for bringing in recombination is its impact on the dimensionality of the genome-wide probing of genetic space. No organism is subject to a single threat at a time. Recombination brings together combinations involving just about every gene in the population. If you focus on a single pairing - say CR's double mutation - you may puzzle as to why it hasn't arisen, given that its likelihood is only increased by recombination. But there are thousands of genes in the genome, and novel pairings arise all the time. While you're sitting staring at CR and wondering why it is so rare, thousands of gene combinations are occurring and being subject to the environmental experiment of Natural Selection. Once again, the 'birthday paradox' rears its head. The chance of NO advantageous combinations is exponentially reduced with each additional candidate. And it's that that Behe is trying to argue for from this single instance - CR is rare, therefore this is the limit on evolution. It's a linear extrapolation. But - to the extent that it is a limit, which isn't very much - it is a limit on the single case. Satisfying yourself that multiple genes in multiple genomes in multiple species are subject to this limit is an error in handling probabilities.

      Delete
    35. Allan:

      I'm not persuaded by your arguments. I think what Behe has pointed out is critical, and should be accepted as the starting point for a new understanding of life's diversity. It would be a shame if, with a outpouring of language, the really essential is made blind to the scientific eye.

      Delete
    36. I'm not persuaded by your arguments.

      Heh. I didn't think you would be. But you haven't even addressed them. You might as well just stick your tongue out!

      I think what Behe has pointed out is critical, and should be accepted as the starting point for a new understanding of life's diversity.

      Accepted by whom? It's full of holes, as several have tried to explain, and biologists will not accept it as a starting point for anything until those holes are addressed, not simply wafted aside. There is, as yet, no sound reason to suppose that a single example of a double mutation in an organism with a particular (and somewhat unrepresentative) life cycle can give us any guidance as to the capacity of evolution to explain organic diversity without the import of a Mysterious Tinkerer.

      Delete
    37. Lino, like many creationists, attempts to wear us down by sheer repetition. Behe said that protein binding sites can never evolve. We have counted numerous examples of protein binding sites evolving-- in Vpu in HIV and Vpr in HIV-2, snake venoms, affinity maturation in the immune system, bovine seminal ribonuclease, etc. etc., we piled them up, so Behe was wrong although IDcreationists change the subject when we list counter-examples.

      Behe said that malaria CR required two simultaneous mutations and it would take 10^20 parasites to evolve CR, and that this was applicable to all protein binding sites. The first two mutations in CR were sequential, not simultaneous, so Behe was wrong. It does not take anywhere near 10^20 organisms to evolve those two mutations, so Behe was wrong by a factor of about 10^5 according to my calculations which are more detailed than Behe's or Larry's. Behe never gave any reason why all protein binding sites should work like CR, and we know experimentally that many evolved and still evolve more quickly with fewer changes.

      Behe was proven wrong, experimentally and mathematically, on every every point. The IDcreationist response is to keep repeating and repeating falsehoods in the hopes of exhausting us.

      Delete
    38. Diogenes:

      Hilariously, Lino invokes the "Why don't chimps give birth to humans?" argument, one of the dumbest in all the arsenal of creationism:

      You've completely missed the point I was making. You have things backwards.

      Delete
    39. Allan:

      Heh. I didn't think you would be. But you haven't even addressed them. You might as well just stick your tongue out!

      Funny, Allan, from my perspective it looks like all you've done is stick your fingers into your ears.

      Delete
  10. So according to standard ID "theory", when things evolve easily, it's just "microevolution" with all the usual ID talking points and caveats (no new information/ no new function/loss of function/mutations only degrade information etc. etc.)

    .. and when things evolve infrequently/take more time to evolve, it's not evolution at all? Or what?

    There's a funny disconnect in all of this. We will often hear IDcreationists complain that we can't show them macroevolutionary change in their lifetimes (at least, on a scale that would satisfy them). Why haven't we been able to show them a fish turning into an amphibian, for example? The usual answer is that it would take a lot of time of course, probably hundreds of thousands of years at least, for the right mutations to accumulate and so on. The IDcreationists are of course not satisfied with this and usually respond that this is just an ad-hoc excuse, which they will caricature with the saying "if nothing happens, just add millions of years of waiting" and similar silly nonsense.

    Now we have an example of something that evolves slowly because it requires multiple sequential mutations, some of which are neutral or slightly deleterious. Nevertheless, It has evolved over ten times. But to the IDcreationists, this is somehow evidence against evolution? How? What am I missing here? Are the IDiots trying to argue that evolution simply cannot be slow and that some things, because they are improbable and therefore take more time to evolve, somehow actually do not evolve?
    That the ten independent, detected origins of chloroquine resistance wasn't due to evolution, but ... magic? What is the purpose of Behe's argument in the end? What is he trying to establish by babbling about how rarely chloroquine resistance evolves? If evolution is slow/takes a long time, it's not evolution at all? That if some things would take even longer to evolve than chloroquine resistance, they could not possibly evolve at all?

    How does that follow? Why do they think there's 5 to 20 million year timespans involved in the major transitions(like fish-tetrapod, dinosaur-bird, terrestrial mammal-whale etc.) in the fossil record? Is their designer picking out mutations so slowly it takes millions of years for him to achieve the desired effect?

    ReplyDelete
  11. Ultimately, the question Behe is trying to ask is whether or not it is reasonable to believe that the great diversity of life we see can be produced by natural means when advantageous mutations are so unlikely. To get a new species you need a sequence of mutations leading from one species to another where each mutation has a survival advantage and is also probable enough to get you to a new species in the time available. The point is Behe wants to see someone who believes in natural evolution present some numbers rather than a "just-so-story" to back up their theory of evolution.

    ReplyDelete
    Replies
    1. To get a new species you need a sequence of mutations leading from one species to another where each mutation has a survival advantage and is also probable enough to get you to a new species in the time available.

      Please try reading slowly for comprehension. In particular, pay attention to mentions of genetic drift, neutral theory, and probabilities involving steps where there are deleterious mutations.

      Delete
    2. Anonymous: " The point is Behe wants to see someone who believes in natural evolution present some numbers rather than a "just-so-story" to back up their theory of evolution."

      We have presented numbers. I did. But every single anti-evolutionist does not tell the truth about what we calculated. I did a calculation for more realistic and detailed than Be he's and I showed he was wrong by a factor of ~10,000. Behe described my calculation, but he did not tell the truth about what I computed.

      Delete
    3. Behe is trying to pick the mote out of your eye while he has a log in his, as he writes here in response:

      "again not all that far on a log scale. Either or both of these values can easily be reconciled to White's calculation of 1 in 1020 by tweaking selection coefficients or by inferring that a further mutation is needed for effective chloroquine resistance in the wild, as Professor Moran noted."

      Yes, not that far on a log scale! And you can always tweak or twerk the coefficients.

      Delete
  12. Remember how chloroquine resistance required 4 mutations in total, 2 of which were neutral if not slightly deleterious?

    If something that requires 4 sequential mutations (some of which aren't even beneficial) takes a decade to evolve in a population of 10^12 individuals, how much evolution by point mutations only, is possible in 3.5 billion years, in a total population of about 10^30 cells(a rough estimate of the total number of cells on earth)?

    Well, 10^30 is 10^18 times larger than 10^12.

    3.5 x 10^9 years / 10 years = 3.5 x 10^8 total decades, in each of which 4 mutations evolve.

    How many total mutations have evolved then?
    10^18 * 3.5x10^8 * 4 = 1.4 x 10^27 total mutations.

    Shit, looks like more than enough to evolve basically every protein and regulation site known to man, and then some, during "the available time". And we haven't even factored in things like gene duplications, frameshifts, polyploidy and so on.

    What Behe says he wants to see but is never done, population geneticists have been doing for over half a century. He just doesn't like the implications, so he's sitting back on a friendly website bamboozling his deluded audience who's all too keen on believing his assertions.
    For example, the idea that nobody has bothered doing calculations such as these. We know the IDiots don't bother checking when a fat fucking LIAR like Behe makes a claim like that, so the issue is he can get away with it. And what do we get here now? - IDiots regurgitating a lie they didn't bother checking.

    Next time one of the higher-ups in the ID movement makes assertions about what "Darwinists" haven't bothered doing, go check if that claim is actually true first. You might be surprised.

    It's funny because, in Stephen Meyers "Darwin's Doubt" screed, Meyer cites Susumo Ohno, a famous population geneticist, in support of a claim of Meyers(which is unrelated to the subject here).
    But in the actual paper that Meyer references, Susumo Ohno does exactly the kind of calculation that Behe says evolutionists don't do - about how many mutations would be required and how long it would take and so on, using accepted substitution rates, to explain the cambrian diversification.
    These are in fact the kinds of calculations that support evo-devo and evolution primarily by rewiring gene-regulation, which is also Ohno's conclusion.

    Why didn't Meyer tell Behe that he's found a paper by Ohno containing what Behe claims never is done? And it's not like Ohno's calculation is the first or only of such calculations done by any stretch.

    That's because both of them are working for a religious propaganda mill, and they're counting on their keen-to-believe audience not actually bothering to verify their claims and read their references.

    It's a fucking disgrace.

    ReplyDelete
  13. Diogenes:

    Maybe I'm wrong, but it seems clear to me that we're looking at two distinct events, asexual reproduction within the human host followed by sexual reproduction within the mosquito, each of which has its own probabilities.

    The probability of a gametocyte ingested by the mosquito from a human host, and which will then go on to form an 'oocyst,' is 1 in 10^12. And then the probability of one of the sporozoites produced within the 'oocyst' entering another human host is 1 in 10^3. Therefore, the probability of a sporozoite entering a host that is CR is 1 in 10^15 since we're dealing with 'independent' events.

    IOW, within the 'oocyst' we have 1,000 different copies of the 'one' genome that entered. It's possible that during the growth of the population within the 'oocyst' that the initial genome ingested by the mosquito, which was CR, then 'loses' the needed mutation during one of its replications. There are a 1,000 chances of losing it. It seems to me the only way of accounting for this possibility is 'mulitply' and not to 'add' probabilities.

    E.g., what if the oocyst made 10^8 copies? Then the probability would be likely that one of the CR mutations would be lost. But it's not 10^8, but 10^3. Nevertheless, there's a chance the CR could be lost within the mosquito. The possibility of mutation continues within the oocyst, and that chance is represented by the 10^3 factor.

    Let's put it this way: what if there were only 1,000 copies of the malarial genome found within a human? And then let's say that it took 100 million infections before CR developed. What would be the probability of CR developing? Wouldn't it be 1 in (10^3 replications per person)(10^8 persons)? I don't know how to present this more clearly.

    Again, maybe I'm looking at this wrong, but I don't think so. Everything hinges on whether you consider the second probability dependent on the first probability. As I see it, what happens in the human host is independent of what happens in the mosquito.

    BTW, the word 'stupid' is not helpful; it's simply antagonizing.

    ReplyDelete
    Replies
    1. There are a 1,000 chances of losing it.

      There are lots of chances of losing it - recovery, death, failure to be bitten again, or to be in the sample drawn by the biter. Loss to reversion to wild-type is negligible though - if there are more cells without it, and mutation rate in both directions is the same, each loss will be balanced by many more gains. It's no good looking to mutation to save you from the implications of mutation! But either way, of course it must be produced multiple times in order to get one out instance of a single carrier.

      Delete
    2. Lito...."The probability of a gametocyte ingested by the mosquito from a human host, and which will then go on to form an 'oocyst,' is 1 in 10^12. And then the probability of one of the sporozoites produced within the 'oocyst' entering another human host is 1 in 10^3."

      How did you arrive at these numbers?

      Delete
  14. ha, i love this shit.

    behe's got diogenes flappin' at the gums, calling him a LIAR. Didn't diogenes realize the moment he opened his mouth he was beat?

    Oh, behe didnt consider this, oh,oh, he didnt consider that. Oh, that LIAR Behe!!!

    F#$%^, the fact that he's got you jackin' at the jaw and Moran posting several post on Behe and the continuous prattle about how nobody but Moran and friends understands evolution says all we need to know: evolution is an exclusive club. you just can't coming walking in, asking for a drink? There' this process, this lengthy application process, thank you.

    diogenes: "LIAR, LIAR, pant on fire"

    The rhetoric is outstanding, just outstanding!!!

    ReplyDelete
    Replies
    1. nobody but Moran and friends understands evolution says all we need to know: evolution is an exclusive club

      Not exclusive at all, all you gotta do is understand some pretty simple math. Here, I'll show you:

      You do understand, don't you, that even though the PowerBall lottery is held twice a week, and the odds against *you* buying the winning ticket are 175 million to one, it won't take 87.5 million weeks until *someone* buys a winning ticket? Right, you now understand the math of evolution better than Dr. Behe.

      What Dr. Behe has been doing is exactly the same as reading in the papers every couple of weeks that someone won the PowerBall, while continuing to insist there can only be one winner every 87.5 million weeks. It's really just that simple, and Dr. Behe is really just that wrong.

      Delete
    2. "There' this process, this lengthy application process, thank you."

      Indeed. It's known in many circles as an education.

      Delete
    3. So Paul McBride, Behe doesn't have an education, James Shapiro doesn't have an education, Jonathan Wells doesn't have an education, Scott Minnoch doesn't have an education, etc, etc, etc.

      A scientist could have a F#$%n' slew of education, but Paul McBride just couldn't concede that they were educated. When push came to shove, he would blurt ' he's not qualified to comment because he's not an evolutionary biologist (the favorite firewall comeback), he's over-qualified, he's under-funded, he's over-enthusiastic for his position, under-estimating of evolution's capabilities; anything but concede evolution has issues, major issues.

      Delete
    4. now judmarc, the fact that someone wins the powerball, never the same individual twice, and virtually if ever, two powerball wins in two draws suggests that Powerball was designed with this knowledge in mind; or Powerball as we know it would be doomed.

      IIANM, Powerball makes beaucoup bucks because it was designed that way.

      Evolution required an endless string of Powerball wins. Intelligence makes that happen. It follow then that......

      Evolution is a designed program. Does not need external tickering.

      A further education (HT to McBride) will elucidate the gaps in our meager understanding of the designed in command and control parameters involved in the initial execution. further development and subsequent maintenance processes involved in evolution.

      IANS, it is Behe who is calculating in the real world.

      Delete
    5. Steve, you are totally wrong. There is no credible evidence evolution is a "designed" program. Nothing we know about the natural world requires this. Behe is not calculating in th real worls, he is calculating in his world, where he makes all the assumptions and all the rules,

      Delete
    6. Evolution required an endless string of Powerball wins. Intelligence makes that happen.

      Only if you think the world you see around you is the world that had to be.

      Where are you sitting right now and what are you doing? What do you think are the chances that you would be where you are right now if calculated at your birth? As close to zero as possible without being zero, yet here you are!
      Or did god micromanage all of the millions of incremental circumstances in your life to get you to this supposed necessary place in the here and now?

      Delete
    7. Steve: "two powerball wins in two draws suggests that Powerball was designed with this knowledge in mind"

      Nothing analogous to this has ever happened in the history of evolution.

      "Evolution required an endless string of Powerball wins. Intelligence makes that happen."

      What a non sequitur! Powerball wins happen once a week, and it's due to sheer randomness. Maybe Steve had something else in mind, but if he's going to blurt analogies he should think up better ones.

      Delete
    8. Hey Steve,

      What's next, images of the virgin mary in random oil slicks and grilled cheese sandwiches and jesus in dog's assholes ?

      Delete
    9. Steve, an education in a subject is not the same as learning about it for the express purpose of attacking it. Jonathan Wells and his writings on junk DNA are the perfect example.

      Delete
    10. Case in point:

      Wells said that "destroying Darwinism" was his motive for studying Christian theology at Yale and going on to seek his second PhD at Berkeley, studying biology and in particular embryology:


      Father's [Rev. Moon's] words, my studies, and my prayers convinced me that I should devote my life to destroying Darwinism, just as many of my fellow Unificationists had already devoted their lives to destroying Marxism. When Father chose me (along with about a dozen other seminary graduates) to enter a Ph.D. program in 1978, I welcomed the opportunity to prepare myself for battle.

      Delete
    11. It's a good plot for an academic spy thriller: "The mole that came out of the hole: The wasted life of Jonathan Wells"

      Delete
    12. Not sure about the spy thriller, at some level you have to be able to sympathize with the protagonist and that would be quite a stretch with Wells.

      Delete
  15. Much of this discussion doesn't focus on what's important. Yes, there are a lot of mistakes in Behe's argument, but the general direction of it is interesting. We can grant Behe that a mutation might occur with a rate of 1E-10, even if he estimates this in the wrong way. Yes, Behe's implicit assumption that there is only one way to generate resistance is illegitimate, but we can still use this as an example.

    Behe’s calculation includes a number of mistakes, but we can ignore all but two of them.

    The larger of the two problems is that Behe makes an estimate for the rate of evolution of 1 complex adaptation in mosquito, and jumps from that to a conclusion about ANY complex adaptation in humans. This is like assuming that over the past million years, the only possible way for humans to adapt by a double mutation would be a specific pair of mutations like the ones in chloroquine resistance.

    That's not the right way to do it. Once we know how many possible mutations there are, we'll want to take that number and calculate the number of unique pairs, which is X(X-1)/2. Then we'll need to make a wild guess about what fraction are beneficial. I'm going to guess 1 in 1000 (but I'll also calculate assuming 1 in a million).

    Now, let's work the numbers.

    In the human genome, there are 2E9 bp, and for each one we could count 7 point mutations (3 substitutions, 1 deletion, 4 insertions-- of which 1/4 are redundant). There are lots of other types of mutations, and the numbers get really huge. The number of sequences of length i is 4^i, thus the number of possible de novo insertions is 4^i * n, where n is the genome size, which is ~ 1E6n for i = 10. A translocation has a point of origin chosen from the length of the genome, it extends up to r bp, and has a point of insertion chosen from the genome, so the count of translocations scales with rn^2! That's insanely large! Maybe we don't want to count these if we think they occur at rates much less than 1E-10 each. But we need to include at least 1.4E10 point mutations.

    The number of double-mutant combinations is then 1.4E10 * (1.4E10 - 1)/2 ~ 1E20. If we assume that 1 in a thousand combinations is beneficial, then we have to count 1E17 (and if we assume 1 in a million, that's 1E14).

    Thus, we have 1E17 (or 1E14) *beneficial* combos each with a rate of 1E-20, so the total rate per person generation is 1E-3 or 1E-6. In other words, you might have already met someone with a beneficial double mutation. This may seem counterintuitive, but look at it this way. Humans have a mutation rate such that the total genomic rate is about 1 (1.6 according to Drake, et al 1998). There are 0s, 1s, 2s etc in that distribution. You have some double-mutants in your family already. We just assumed that 1 in 1000 (or 1 in a million) are beneficial, so voila! multiple beneficial double mutants just in NYC.

    What about all of human evolution? If humans had a population size of a million for the past million years, that is ~1E11 person-generations, as a rough order-of-magnitude estimate. That's surely an over-estimate. With these numbers, humans have experienced many beneficial double mutations. Even if we assume that only 1 in a million doubles are beneficial, we still get about 1E5 of them in the course of human evolution.

    That solves the biggest problem with Behe's argument. The next problem, still quite significant, is the lack of population genetics. This post is long so I'll address that separately

    ReplyDelete
  16. (continuing)
    Of those many beneficial double mutations that have happened in human history, most got lost. We have to take population genetics into account. When we do that, we don't want to frame this in terms of a double-mutation, as if we have to wait for two mutations to happen in one individual. Actually, a mutation could happen in an individual, and the second in the individual's grand-child, even if the first mutation is deleterious (because low-fitness individuals have children and grand-children, just not quite as many as high-fitness ones). If we extend out this line of thinking, the probabilities start to add up, and we discover an important population-genetic phenomenon of "stochastic tunneling", where we can get from 0 --> 1 --> 2, even when 1 is deleterious and never takes over the population.

    And I know this because population geneticists have already looked at this question in some ways, and it is interesting. It is reasonable to ask how adaptations that require multiple mutations can emerge in evolution, and whether we can calculate any limits to that. I think the most general work is Lynch's paper, which cites Behe (http://mbe.oxfordjournals.org/content/27/6/1404.full.pdf+html ).

    So, we want to reframe this problem in terms of what Lynch calls "complex adaptations," i.e., adaptations that depend on multiple mutations, without restricting it to waiting for a double mutation. Complex adaptations are going to be more likely with more time, and less obviously, with larger population sizes (see Lynch). Humans are at a disadvantage relative to mosquitos and fruitflies and E. coli. At the end of the day, there are going to be some kinds of complex adaptations that are likely in fruitflies but negligible in humans.

    And basic models of this are available in Lynch's paper. This covers the case where the mutations are beneficial together and deleterious (or neutral) separately, and it also considers mutator alleles, and > 2 mutations.

    ReplyDelete