Monday, March 15, 2010

Predictions of Intelligent Design Creationism

I love it when they make predictions. Here's what Casey Luskin just posted over on Evoluton News & Views (sic) [A Response to Questions from a Biology Teacher: How Do We Test Intelligent Design?]

Regarding testability, ID makes the following testable predictions:

(1) Natural structures will be found that contain many parts arranged in intricate patterns that perform a specific function (e.g. complex and specified information).
Such natural structures exist. They are perfectly compatible with evolution. This prediction does not distinguish between Intelligent Design Creationism and real science.
(2) Forms containing large amounts of novel information will appear in the fossil record suddenly and without similar precursors.
The vast majority of of species with novel information have arisen gradually with plenty of transitional fossils to document their evolution. In a small number of cases (e.g. Cambrian Explosion) the evidence for evolution from known precursors is too sophisticated for creationists to follow. This prediction is not supported; therefore, Intelligent Design Creationism is falsified for all species where we have lots of data.
(3) Convergence will occur routinely. That is, genes and other functional parts will be re-used in different and unrelated organisms.
Convergence is perfectly consistent with evolution. This prediction is meaningless.
(4) Much so-called “junk DNA” will turn out to perform valuable functions.
Most of our genome is junk. The evidence for this is overwhelming. Intelligent Design Creationism has been falsified.

That was easy.
In this regard, ID is falsifiable. When we test these predictions, ID passes those tests.
No it doesn't. Two of the "tests" are consistent with evolution so the "test" is meaningless. Intelligent Design Creationism fails the other two tests.


26 comments :

  1. Why is it that the ID field has never presented a detailed phyologenetic tree that specifies ID events in the evolution of life? After all, Behe and Meyer claim ID does not refute common origins and evolution, only that they negate out any possibility of an Intelligent Designer.

    Here is a hypothetical textbook presenting a side by side comparison of what Evolution and Intelligent Design actually predict in regards to the development of life:
    http://i.imgur.com/25MC5.png

    ReplyDelete
  2. How is number 4 even a test? All he is doing is waiting to see if something turns up.

    ReplyDelete
  3. Actually Nosmo-King, a test doesnt have to be currently testable or tested. For example, there are specific predictions (tests) of the kind of fossils that should be found at certain strata (Read Your Inner Fish for one good example). Waiting to see, is ok in science. However, in regards to #4 a prediction is also that loss of this junk DNA would be detrimental, since it does something. However, data in mice suggests that isn't the case.

    ReplyDelete
  4. Have you noticed that the fish in that image is actually a whale?
    I see several other mistakes, but I let you to find them.

    ReplyDelete
  5. Malena,
    The image is intended only to convey the lack of detail on the side of ID side. Then again, the beauty of a scientific model is that it can be verified or challenged with more scientific evidence, as your comment demonstrates.

    In comparison, Stephen Meyer's promotion of ID has no such detail. While Meyer says he does not refute speciation per se., he does claim that new and different body plans required an Intelligent Designer.

    Now a biologist might assume he is using the term body plan in the traditional sense, as in: The Chordate's body plan is defined by the notochord. It's this type of feature that Evolutionary theory predicts Chordates like Sea Squirts, Chickens, and Mankind should share if they do indeed share a common ancestor.

    And this is where ID is both conceptually barren and purposely deceitful; the ID theorist never specifies what they mean by Species or Body Plan or Common Descent in relation to a phylogenetic tree. Is it any wonder then that ID lacks an explanation - visual or otherwise - to account for the appearance of Vertebrates, Mammals, or Mankind?

    ReplyDelete
  6. Convergence is perfectly consistent with evolution.

    True. But convergence of the entire 3D fold for a moderately-sized protein is incredibly unlikely.

    ReplyDelete
  7. Casey sure uses a lot of weasel words in his scientifical predictions.

    ReplyDelete
  8. Ah I see Anonymous has made my point, albeit with a lot more eloquence and insight. :D

    ReplyDelete
  9. DK you said

    True. But convergence of the entire 3D fold for a moderately-sized protein is incredibly unlikely.

    actually, there is remarkable convergence in protein tertiary structure even in some cases where the genetic codes are disparate

    ReplyDelete
  10. sorry, should read there can be remarkable convergence

    ReplyDelete
  11. How is number 4 even a test? All he is doing is waiting to see if something turns up.

    Well it is a prediction, at least. Too bad we don't know what Casey means by "much" though! Oh well. I predict that much of tomorrow will have much much earthquakes on much of the planet.

    ReplyDelete
  12. Might as well throw in a number 5.

    5) Intelligent Design predicts there will be earthquakes tomorrow. In this regard, ID is falsifiable. When we test these predictions, ID passes those tests.

    Makes about as much sense.

    Might as well throw in some outright lies while we're at it... Darwinist poopy-heads don't like earthquakes, but Intelligent Design makes people more earthquake savvy.

    There you go. You got some "big whoopity doo-daw" mixed in with some bald-faced lies. You are now officially a genu-ine bone-fide IDiot theorist.

    ReplyDelete
  13. In addition to other problems, #1 and #3 aren't even predictions, since both are already known to be the case.

    For his next amazing feat, Casey will use ID creationist insights to predict the Earth will be found to be round not flat.

    ReplyDelete
  14. Point 1:
    Bzzzt. ID predicts no such thing. ID says that only intelligence CAN produce these things, not that it WILL.

    Point 2:
    Bzzzt. Creationism predicts this, perhaps. ID doesn't. Why? Creationism (the YEC variety) says that the designer simply creates things willy-nilly. ID says no such thing. Unless ID=creationism, that is... Casey scores an own goal.

    Point 3:
    Bzzzt. This is even worse than point 2. Neither creationism nor ID has anything to say in these matters.

    Point 4:
    Bzzzt. Perhaps creationism with it's perfect designer predicts this. ID doesn't. Unless ID=creationism... Another own goal.

    Note: Above, I'm evaluating ID's predictions scientifically (where ID predicts nothing), not politically (where it predicts everything creationism does).

    Of course, since ID doesn't scientifically predict anything at all, it isn't science.

    ReplyDelete
  15. Of course ID doesn't make any of those "predictions". The ID hypothesis tells us nothing about the capabilities or intentions of the designer, so he could have done anything at all. For example, he could have included junk DNA with no function at all.

    Typical ID drivel.

    ReplyDelete
  16. Regarding #1 & #3: I predict that a lot of sand will be found on Miami Beach. Can I get my official ID "scientist" merit badge now?

    ReplyDelete
  17. actually, there is remarkable convergence in protein tertiary structure even in some cases where the genetic codes are disparate

    I'm not entirely sure what you mean by this. The first question is whether you actually meant "genetic sequences" rather than "genetic codes," since the genetic codes of all known life forms on Earth are remarkably similar, which is evidence for descent from a common origin.

    ReplyDelete
  18. actually, there is remarkable convergence in protein tertiary structure even in some cases

    Actually, there isn't. Known and accepted examples of convergence are small substructures and/or active sites. The rest of the similarities in 3D folds is generally viewed as an evidence for divergence - precisely because probabiliy of convergence is so low. Although some, like Larry, disagree with this POV.

    ReplyDelete
  19. Dk says,

    The rest of the similarities in 3D folds is generally viewed as an evidence for divergence - precisely because probability of convergence is so low.

    How do you know that the probability of convergence is so low? How did you do your calculations?

    ReplyDelete
  20. This is something (in numbers 1 and 3) that I see in other forms of pseudoscience and fringe science; he's taking observations of things that are known to be true, and calling them predictions. I call them "Sylvia Browne predictions".

    ReplyDelete
  21. precisely because probability of convergence is so low.

    How do you know that the probability of convergence is so low? How did you do your calculations?


    I am just a glorified technician - I do almost no calculations :-). What I wrote was based on my understanding of the problem ("gut feeling" if you wish) and cursory reading of the relevant literature (cursory because they all seem to be saying the same thing and my gut feeling says it is self-evident). In this respect, I should probably take back "precisely because" part above. Few things to note:

    1. We don't have structures that are transitory between two folds. If convergence is common/possible, there would have to be some.
    2. We have plenty of examples of convergence of the function. Nature does not optimize, it simply builds upon existing - whenever it starts from combinations of elements, it is easy to converge upon the same function but the underlying basic structure is always different. Works the same way for proteins and organs.
    3. There are very few folds. Orders of magnitude less than theoretically possible. That is hard to explain if convergence is happening. Even in transmembrane proteins (where the physical requirements for folds are relaxed) the variety of folds is no bigger than for soluble proteins.
    4. Randomly picking up very limited similarity in 1D (sequence) is a lot more likely than randomly matching remote similarity in 3D (fold). This is simply because 3D space is much bigger. Thus, even when sequence signal is degraded, one would expect still a strong signal from a fold. (I do realize it does not prove homology; just that in most cases you should expect it).
    5. E.g., JMB, 2004, 336:695–706:
    "One may thus conclude that structural proteomes of extant organisms are most likely the result of a specific, correlated mechanism of domain evolution, a conclusion that explicitly violates the equilibrium precepts of most convergent models."

    ReplyDelete
  22. DK presents several arguments against the independent convergence or proteins with similar folds.

    1. We don't have structures that are transitory between two folds. If convergence is common/possible, there would have to be some.

    The folds had to evolve from some sequence that did not have all the characteristics of the modern fold. The fact that we don't see modern examples of these primitive "pre-folds" is not really surprising, is it?

    If we don't see examples of transitional molecules for any fold then it shouldn't be a big shock that we wouldn't see such transitional molecules if two random polypeptides each converged on the same stable fold.

    2. We have plenty of examples of convergence of the function. Nature does not optimize, it simply builds upon existing - whenever it starts from combinations of elements, it is easy to converge upon the same function but the underlying basic structure is always different. Works the same way for proteins and organs.

    I don't understand this argument. There are lots of beta-barrel enzymes and proteins that have different functions and the same fold. They are presumed to be good examples of convergence. What's wrong with that logic?

    3. There are very few folds. Orders of magnitude less than theoretically possible. That is hard to explain if convergence is happening.

    No it's not. You've got the argument completely reversed. If there's a relatively small number of stable folds then it's not surprising that several different primitive proteins converged on one of these folds. There were no other options for survival.

    Even in transmembrane proteins (where the physical requirements for folds are relaxed) the variety of folds is no bigger than for soluble proteins.

    If there's only a small number of folds that are compatible with being in a membrane then why wouldn't several different proteins evolve that fold?

    4. Randomly picking up very limited similarity in 1D (sequence) is a lot more likely than randomly matching remote similarity in 3D (fold).

    I don't think anyone is arguing that the folds arose randomly. I assume they evolved under selection for stability.

    Do you think they arose spontaneously in one shot?

    This is simply because 3D space is much bigger. Thus, even when sequence signal is degraded, one would expect still a strong signal from a fold. (I do realize it does not prove homology; just that in most cases you should expect it).

    I don't think that stable 3D space is nearly as big as you do.

    Your argument requires that two proteins descend from a common ancestor but sequence similarity will be completely lost while preserving protein structure. That doesn't seem very likely to me.

    5. E.g., JMB, 2004, 336:695–706:
    "One may thus conclude that structural proteomes of extant organisms are most likely the result of a specific, correlated mechanism of domain evolution, a conclusion that explicitly violates the equilibrium precepts of most convergent models."


    I'm glad you recognize that there's a legitimate scientific debate going on over whether convergence or divergence is the best explanation of the data.

    At this point, it would be wrong to assume that all proteins with similar folds MUST be homologous. That may be your opinion but it's not a scientific fact.

    ReplyDelete
  23. The folds had to evolve from some sequence that did not have all the characteristics of the modern fold. The fact that we don't see modern examples of these primitive "pre-folds" is not really surprising, is it?

    No, of course it's not. The "pre-folds" were small folded elements with limited untility and stability unless part of a larger fold. So your scenario would envision independent discoveries of the same fold because the stable 3D space is so small. (Certainly we don't have examples of TIM barrel morphing into beta propeller, so I hope you discount at least this possibility!)

    The idea that there are only so few stable folds possible strikes me as ludicrous. For ~ 25K protein a number of different folds is >20^200. OK, take away neutral substitutions you still have something like 20^100. And you are saying that out if this essentially infinite number only 5,000 combinations are stable??? Considering that all a fold needs is to hide enough fat inside, that's not very likely. I've seen couple theoretical works explicitly modeling this and they concluded that based on current understanding of protein stability there should be many, many more folds (one paper aptly made an analogy to a "dark matter" problem).

    If there's only a small number of folds that are compatible with being in a membrane then why wouldn't several different proteins evolve that fold?

    But there should be a lot more folds allowable for intramembrane space! Somehow they are not realizing (and that's because there is no need - few original hits were plenty enough).

    There are lots of beta-barrel enzymes and proteins that have different functions and the same fold. They are presumed to be good examples of convergence. What's wrong with that logic?

    Examples, please? Are you talking about TIM barrels? I'd argue that they are all homologs.

    ReplyDelete
  24. Number four has just been proven to be true. Almost 80% of "junk" DNA has a purpose. I do agree with the rest, though.

    ReplyDelete
    Replies
    1. No it hasn't, the ENCODE project's sensationalistic claims have been thoroughly refuted by the scientific community. Heck, even the authors of the ENCODE claims have come out and retracted some of their more sensationalistic claims (among them the 80% functionality claim you speak of).

      Delete