|
Author
|
Topic: Behe IC test
|
zygotecowboy
Member
Member # 220
|
posted 02. April 2002 13:41
Michael Behe defines irreducible complexity (IC) as the following:
quote:
By irreducibly complex I mean a single system which is composed of several well-matched, interacting parts that contribute to the basic function, and where the removal of any one of the parts causes the system to effectively cease functioning. (Behe 1996, DBB: 39)
Behe suggests that a mousetrap fits the bill as an IC system. Since I was unwilling to accept this claim by assertion alone, I decided to do the IC test myself. I headed to the local hardware store and bought 5 mousetraps. I removed one of the “several well-matched, interacting parts” from each of the traps. Sure enough, the traps no longer worked. One would no longer snap because the spring is gone. One simply fell to pieces because the base was not there to hold the other pieces together. Another no longer would hold down the kill-bar in a retracted position. You get the idea...
My test indicates that “the removal of any one of the parts causes the system [mousetrap] to effectively cease functioning.” I removed all the parts one by one, and the mousetrap no longer works. I did the test for IC, and I am convinced that mousetraps exhibit irreducible complexity.
Much has been made about IC systems in a biological context. The current empirical crux of intelligent design theory rests on the existence of biological IC systems. For instance, Behe has suggested that the blood clotting cascade, the immune system, and the bacterial flagellum are examples of IC systems. However, has anyone done the IC test for these systems? Has anyone removed each and every part one-by-one and demonstrated the system has ceased functioning?
I’ve posed these questions before in another forum, and Mike Gene was gracious enough to provide me with his interpretation of the situation. He suggested that the IC state of the flagellum was demonstrated by gene knockout experiments. However, I contend that this is not an adequate IC test. First of all, many (most? all?) proteins that make up the bacterial flagellum are duplicated multiple times within any single flagellum. Dozens of nexins, dyneins, nine subfiberA’s, subfiberB’s, radial spokes, and two central singlets as well as at least forty other different types of proteins make up or are involved in the production of only one flagellum. Knocking out a gene for any of these proteins would eliminate all of the multiple copies used. Certainly this would cause the flagellum to “effectively cease functioning”. However, an analogy should illustrate that this is not an appropriate IC test, at least according to how Behe describes it.
Suppose we perform an IC test on a car. Would the proper test be to remove all the O-rings from all parts of the car at the same time? Or all the tires? Or all the nuts? Or all the bolts? This is not what is called for in Behe’s description of IC. He says that “the removal of any one of the parts causes the system to effectively cease functioning.” A proper IC test would be to remove each bolt one-by-one. Each bolt is an individual part playing its own individual role in the overall function of the car. Every single bolt summed together make up a class of parts.
In the bacterial flagellum, each dynein is an individual part, each nexin is an individual part, and each radial spoke is an individual part. Knocking out a gene for a protein product is not removing “any one of the parts”, it is the removal of an entire class of parts. This is akin to removing all the bolts from a car at one time. This is not what Behe calls for in his definition of IC, at least the version of IC Behe gives in DBB.
In addition, genetic knockouts remove the class of parts even before the “machine” is assembled; there is no “removal” as Behe’s definition calls for since there is nothing to remove in the first place. This seems to run contrary to Behe’s notion of IC.
I suppose it may come down to a quibble of what the definition of “parts” is. Is every protein an individual part that contributes to the basic function? Or do we need to think of flagella as rotors, drive shafts, hooks, etc. Nevertheless, wouldn’t the demonstration of removing each and every protein one by one causing a flagellum to cease functioning be a much stronger case for IC than simple gene knockouts?
Have those involved in formulating an ID research program made plans to explore this area of research? Perhaps a discussion of the techniques required for a proper Behe IC test would be best. Do techniques exist that can manipulate complete biological systems on the molecular level, and remove individual proteins (parts)?
zc
IP: Logged
|
|
Nelson Alonso
Member
Member # 52
|
posted 02. April 2002 14:12
Zygote:
In the bacterial flagellum, each dynein is an individual part, each nexin is an individual part, and each radial spoke is an individual part. Knocking out a gene for a protein product is not removing “any one of the parts”, it is the removal of an entire class of parts. This is akin to removing all the bolts from a car at one time. This is not what Behe calls for in his definition of IC, at least the version of IC Behe gives in DBB.
Nelson:
I think you are talking about the eukaryotic flagellum, not the bacterial flagellum. Genetic analysis, which is generating mutations and selecting those that are relevant to or have an effect on a cellular function, in determining the number of genes involved in a structure oriented system, and how these genes or gene products interact to produce function (by analysis of multiple mutants) , is a good method for which the ICness of the system can be teased out. In other words, knock-out experiments,is, at its core, what IC is. Dynein, for example, can be said to be a part. Remove the motor, and the system does not work. I agree that the closer we look, the more complexity , and perhaps, we will even see more irreducible complexity. Although "part" here may not precise, it is still very useful in the way that Behe has used it.
[ 02 April 2002, 14:13: Message edited by: Nelson Alonso ]
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 03. April 2002 00:40
Zygote,
Interesting post. This is right off the top of my head, but I want to take a stab at answering your question.
First, the gene seems to be the basic unit that evolution can work with. That is, knockout experiments get rid of all gene products that are made by a gene, but this seems like the reverse of what evolution would have to do. Evolution deals with genes as units, not with individual gene products, so no matter how many proteins are made by a gene, they are going to have to be either all present or all absent in the evolutionary process.
Second, your analogy with cars and O-rings is a bit inaccurate. The reason is that the gene products that are duplicated in, say the bacterial flagellum or eukaryotic cillium (which you seemed to be talking about in your post) are, as far as I know, identical to each other, both in structure and in function. So it would be like someone taking an engine and removing, say, all the piston rings. Now, certainly this engine won't work, and it's fair to say that the piston rings are part of the irreducibly complex system, even if the engine is capable of functioning with one of them missing. The reason is that there is a functional threshold that requires at least X number of piston rings. The function of piston rings is absolutely essential to the engine's function. If we remove two rings, the engine probably won't work, and if we remove them all it definitely won't work. There is a basic functional threshold the system has to cross to achieve functionality. In other words, the duplication of the rings is necessary because of the increased complexity of the system relative to a simple function that can be flled by only one object (take, for example, the single holding bar of the mousetrap where only one element is enough for function). So my response is that while there may be the possibility of removing one dynein arm, or two, etc., there is some minimum number that are going to be needed to maintain function. This threshold concept actually makes the system more complicated relative to a simpler system that only needs one component, and certainly doesn't in any way suggest that the system is not irreducibly complex (one could even talk about an irreducibly complex core that has extra elements, perhaps as a redundancy for making the system more robust).
Thirdly, your objection seems to get at the question of gene regulation. For a gene to produce the correct number of components, it needs to be tightly integrated with a regulatory network that produces the right amount and also manages to get the gene products to the right place. So far from making the irreducible complexity simpler, we are introducing vast amounts of complexity (probably itself irreducibly complex!)
Fourthly, let's consider the implications of finding that, say, not all dynein arms of a eukaryotic cillium are needed for function. We certainly know that some number of dynein arms are needed for function; the extras are just redundant copies not absolutely necessary. But how can we account for the evolution of redundancies? Once we have the basic function there is no more selection pressure. There is no functional advantage to having a redundant system over having a basic functional system, so no obvious reason why it should be preferred. However, functional redundancy makes a lot of sense from an engineering standpoint and is a common design principle for important systems. Maybe these redundancies are better explained as a result of deliberate design.
Just a few thoughts.
John Bracht
IP: Logged
|
|
zygotecowboy
Member
Member # 220
|
posted 03. April 2002 12:17
quote:
Dynein, for example, can be said to be a part. Remove the motor, and the system does not work. I agree that the closer we look, the more complexity , and perhaps, we will even see more irreducible complexity.
So the dozens and dozens (hundreds?) copies of dynein that are utilized in a single flagellum are a single "part" called the "motor"? Is there anything analogous to this from human technology?
Most human motors are probably irreducibly complex, in that if you remove any single component it will effectively cease functioning. Is this true of the motor of the flagella, with its multiple copies of dynein? Can this be tested with today's technology? If not, I think this is the avenue an ID research program should follow. A proper Behe IC test is paramount to the whole concept of ID, since this is currently the only facet of ID that is based on empiricism.
zc
IP: Logged
|
|
Nelson Alonso
Member
Member # 52
|
posted 03. April 2002 14:47
Zygotecowboy:
So the dozens and dozens (hundreds?) copies of dynein that are utilized in a single flagellum are a single "part" called the "motor"? Is there anything analogous to this from human technology?
Nelson:
Yes for example, one can keep cutting off the corners of the base of the mouse trap and he can have 8+ pieces of wood. All of these can be said to be parts. There is a minimum lenght, however, needed for the base of the mousetrap in order for the mousetrap to work. Again, you are simply pointing out more irreducible complexity. Or take screws, or bolts. If I remove one bolt, I will not have a functionless system, remove all the bolts and the system will collapse. There is, again, a minimum number of bolts, though that cannot be removed, or the system falls apart.
Zygote:
Most human motors are probably irreducibly complex, in that if you remove any single component it will effectively cease functioning. Is this true of the motor of the flagella, with its multiple copies of dynein? Can this be tested with today's technology? If not, I think this is the avenue an ID research program should follow. A proper Behe IC test is paramount to the whole concept of ID, since this is currently the only facet of ID that is based on empiricism.
Nelson:
When you look at the dynein itself as a "part", the fact that we cannot delete each and every single individual piece of the system with today's technology does not render Behe's IC useless. Right now we have a core set of "parts" to work with, it's true, remove one and it doesn't work. It fits the definition nicely.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 04. April 2002 18:56
Zygotecowboy: So the dozens and dozens (hundreds?) copies of dynein that are utilized in a single flagellum are a single "part" called the "motor"? Is there anything analogous to this from human technology?
No, but then again, humans have yet to develop any substantial nanotechnology. When it comes to nanotech, it would be erroneous to think the only difference will be size. Take the bacterial flagellum. Its filament is composed of about 10,000 flagellin proteins. Why would any engineer consider each flagellin monomer as a distinct, separate part? What engineer would design a protein filament with 10,000 completely different proteins?
Or lets consider the FliM rotor of the flagellum. It is built from approximately 35 FliM proteins. This creates something that does not exist in an individual FliM protein - a ring. In the flagellum, FliM does not function as a single protein - it functions as a ring of proteins (in fact, in the cell, most proteins function as part of a complex). Now, the FliM ring is intriguing in that not only is it (and FliG/FliN) rotated by the motor, but it contains the business end for switching. That is, another protein, CheY-P, can interact with the FliM (when phosphorylated) and cause it to reverse its direction of rotation. What is really neat is that there is a built in threshold level of FliM-CheY-P interactions that needs to be crossed to bring about switching. If less than 80% of the FliM ring proteins interact with CheY-P, nothing happens. Then, somewhere between the 85-90% level, there is a extremely rapid increase in the likelihood that switching will occur. We can see the logic of this approach as it dampens the effect of illegitimate interactions between CheY-P and FliM. And it gets even better if we try to switch backwards, as the forward and reverse processes do not superimpose. Once rotating in the other direction (as a consequence of CheY-P interactions), you have to knock out more than 50% of the CheY-P interacting proteins to switch back. In other words, there is no simple threshold-crossing mechanism, but instead, two thresholds are involved. This all suggests that the switch shows hysteresis which, not too surprisingly from an ID perspective, is also employed to provide smooth and adjustable torque control for motor/rotors designed by humans.
On a related point, John writes:
Thirdly, your objection seems to get at the question of gene regulation. For a gene to produce the correct number of components, it needs to be tightly integrated with a regulatory network that produces the right amount and also manages to get the gene products to the right place. So far from making the irreducible complexity simpler, we are introducing vast amounts of complexity (probably itself irreducibly complex!)
Another example from the flagellum. The rod or driveshaft is a helical structure, much like a spiral staircase. Yet it is composed of at least three proteins, flgB, flgC, and flgG, where six, six, and twelve copies are present as part of the structure. It gets interesting when we realize the rod is thought to consist of about 5.5 proteins per complete turn of the helical structure, suggesting we have a complete turn made up of flgB, then flgC, and two complete turns made of up of flgG. Now, there is nothing inherent in the structure of a "spiral staircase" that would order things like this. Thus, there must be some mechanism, extrinsic to the driveshaft, that does the ordering. If this mechanism is essential to the assembly/function of the flagellum, we'd have another good candidate for a built in IC subsystem embedded in the IC flagellum.
In the end, it would be great if we could develop techniques to precisely remove one protein at a time from an in vivo complex. But as Nelson and John argue, this would probably help to reinforce our position, rather than damage it. For gene knockout experiments tell us we need "at least one" copy of the gene product. But clearly, having at least one copy of each gene product that makes up the flagellum, suspended in test tube, is not the same as the functioning flagellum.
[ 04 April 2002, 19:01: Message edited by: Mike Gene ]
IP: Logged
|
|
|
Printer-friendly view of this topic