|
Author
|
Topic: Organisms using GAs vs. Organisms being built by GAs
|
William A. Dembski
Member
Member # 7
|
posted 19. September 2002 18:37
Deanne,
[Okay, no long responses till the weekend.]
Your response shows that scale invariance is dispensable when it comes to understanding the origin of irreducible complexity. Indeed, when I ask for bridge principles, scale invariance disappears and we get the standard coevolutionary story with genes duplicating and then being grist first for neutral and then selective evolution. As for the distinction between scale invariance being a global as opposed to local property, I'm afraid you can't have it both ways, at least not in biology. We are not here dealing with Koch curves that are identical at different scales. Biology gives us specificity at each level in the hierarchy, and that specificity is not identical at different levels of the hierarchy. Thus the criticism I made of scale invariance being a global property still holds. Indeed, that's why scale invariance disappears as soon as you get to your bridge principles. Scale invariance, it seems, functions for you as a nice conceptual framework for keeping the pieces of evolutionary biology organized. Frameworks are all nice and well, but they are not explanatory. They can assist explanation, but in your case the framework is merely shifting the problem rather than resolving it.
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 19. September 2002 18:59
In reply to Paul,
Must...not...spend...all my time...on ISCID...
But: the favoring of diversity in immune system in molecules (and conversely, in the external molecules of diseases) is not a supposition, rather it is a trend we see just about everywhere we look -- the adaptive immune system is just one way to increase diversity. We can tell that positive selection for diversity of immune system genes has been operating for at least the last few millions of years based on substitution rates, and I bet that we'd find this wherever we look. I can literature-bomb you on this later if you like.
For someone with "20+ years in the evolutionary literature" it seems that your questions so far have been pretty easy to answer when you've actually got some reasonably-informed people willing to do it. Papers in the literature don't talk back and aren't specifically designed to meet the peculiar objections of IDists, so perhaps that has been part of your problem. I can give these latest a shot. I will point out where I think you make some fairly elementary errors of understanding RM&NS.
quote:
My original comment/question (2) was poorly phrased. Let me try again. Presumably the proto-RAG transposon could land anywhere in the genome, and be expressed (or not). What evidence shows that the proto-RAG transposon would insert only into a primordial Ig locus, and be expressed only in the cell types where that locus was active?
Sure, the transposon could insert anywhere, just like a mutation could happen anywhere. It does not have to happen miraculously just in that one specific spot. Things like fossil viruses and such have been discovered all over the place in various genomes. The question is whether any of those insertions would ever be selectively advantageous rather than disadvantageous, and there is a good argument that it would be in the specific case of the receptor gene (as we know on independent evidence that diversity in receptors is strongly favored by selection).
This shows a flaw in the understanding of mutation (here including random transposon insertion), i.e. assuming that the changes that we see today were the only ones that happened.
As for why this receptor would only be active in an immune system (lymphocyte-like) cell: organisms had immune system cells before they had an adaptive immune system, and almost by definition an immune system cell will be expressing immune receptors (to get the bad guys) and other cells won't. I expect that this situation is common today but perhaps a resident immunologist can confirm.
As for why RAG expression only happens while the lymphocyte is producing the receptor, it's pretty clear that the proto-RAG protein (which snips the RAG gene out of the genome) will only get it's chance to be expressed when the surrounding receptor is expressed or about to be expressed (that section of the DNA is being unwound, translated into RNA, etc.). Since the process of the RAG protein snipping out the RAG DNA sequence process is somewhat innaccurate, then when this happens in the lymphocyte cell, the descendents of that particular cell will inherit a slightly varied receptor. Other lymphocyte cell lines will have their own slightly different receptors. Voila, a primitive version of the adaptive immune system.
Since the RAG snipping-of-itself-out only occurs when the receptor gene is activated in the lympocyte cell like, the RAG gene persists in the germline and organisms with the RAG gene benefit from the enhanced diversity of receptors.
(from here further modifications, described by Inlay, get us to the modern system, but the major hump is passed)
quote:
2. What evidence shows that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon?
In the first place, all we're talking about is slight variations in the resulting receptor, and we know that cells make use of a diverse array of non-rearranging receptors. In the second place, as noted before these receptor genes are particularly prone to duplication and modification (redundancy is everywhere), all we're talking about here is taking one of these many copies and giving it a slightly different type of mutation. In the third place, if a few of the lymphocyte-like cell lineages had screwed up receptors initially, so what? (Particularly if this proto-lymphocyte was expected to have the "if my receptor finds something I will replicate" quality, rafe or charlie can comment on if cells with non-rearranging receptors tend to have this)
This shows a few more common ID mis-assumptions, namely: (1) there is only one copy to tinker with and that (2) therefore in messing with this one copy you will lose something else. Just ain't so when multiple copies of genes (and near relatives, etc.) are around.
quote:
Or not. Bacteria have nothing resembling the vertebrate immune system, and they're doing just fine.
Bacteria are not large multicellular critters with blood streams and stomachs and such to infect. They do not have to seek-and-destroy
diverse invaders. They are the invaders.
Plus, bacteria have a replication rate at least as fast as that of the immune system cells themselves. The adaptation of the large population of bacteria is (the population's) immune system, but the term has pretty much lost meaning at this point.
Presumably you were trying to say something about taxa without adaptive immune systems. Further considerations that probably favored the original of adaptive immunity in jawed vertebrates as opposed to invertebrates include (1) larger size, (2) longer lifespans and (3) the development of jaws exposing the (increasingly extensive and higher surface-area) gut to foods rather more challenging than our filter-feeding ancestors. Lest you claim that this latter hypothesis is mere hand-waving, it would predict more sophisticated immune systems in the larger, longer-lived, and more predatory lineages across taxa. Octopi, cuttlefish and giant squids might be a place to start looking.
Failure to think carefully about what organisms are and aren't good analogies in testing evolutionary models is yet another flaw we can identify.
quote:
In his talk.design article, Matt Inlay said that the descendants of the original RAG mutant "thrive with the increased immune capabilities." But how does he know this?
I suggest that he doesn't know it. Rather, he's simply waved the wand of "natural selection" over a (putative) historical episode, postulated an entirely hypothetical selective advantage, and called it an explanation.
'Tis is a model that attempts to explain what you guys want explained. The idea that immune receptor diversity is a widely favored trait is indeed testable and indeed is widely confirmed by numerous avenues that we can go into if you like. So it's not "entirely hypothetical".
If you've got your own testable model, though, I'm all ears...
nic
[ 19 September 2002, 19:08: Message edited by: yersinia ]
IP: Logged
|
|
Deanne M. Taylor
Member
Member # 274
|
posted 19. September 2002 21:39
William:
Scale invariance does not disappear on observation. Scale invariance is an intrinsic, mathematical property of a network. It's a mathematical definition. Mathematical definitions cannot disappear on closer scrutiny. They remain, looking back in a sobering stare. ![[Smile]](smile.gif)
A scale-free biological network has qualities which are pervasive, mappable, and measurable whether one is looking into a local or a global behavior of that network. The term 'scale invariant' means the mathematical structure of that network is the same in a local piece of that network as it is all through the rest of the network.
While a biological network is NOT, indeed, a Koch snowflake, its construction possesses the same qualities locally as well as globally, short of a reduction to individual elements. If you descend to single-network components and claim that as the end of the scale-invariance, well, that's reductio ad absurdum, anyway...as you posit IC systems are systems, and not individual components. Individual components do not an IC system make (though one could argue that the protein structure and sequence are, themselves, an ancestral unit of a system with physically smaller components, but we'll go there another time).
So, yes, on your "systems" level, the system will have mathematical characteristics that match those at the local level.
And indeed, the qualities of the network (of the network, mind, not of the individual properties of the pathways that are within that network) are indeed scale invariant. You can make a mathematical measure as to the mathematical qualities of the graph that the network produces and a local piece of the graph will have the same mathematical qualities as the entire graph. When mathematicians and physicists call such graphs/networks "scale invariant" they're using the very definition of that behavior.
I know quite a bit of biology (It's my living at the moment) and I'm well aware that proteins have different functions in different pathways. But that's not what I'm getting at, here. I'm not SAYING that a biological network has to be an ATP synthetase or a kinase cascade or a citric acid cycle all through the cell for it to share local properties. I'm saying that the very mathematical structure of the protein-protein interaction network is scale-invariant AND such scale-invariance, no matter what qualities the nodes have, is a direct result of sequential, preferential addition. That's just mathematics.
My argument has got little to do with the specifics of what the nodes (proteins) do, or how they function. My argument is on the intrinsic quality of the structure of the networks that the proteins make in a living organism. These networks have a mathematical, scale-invariant quality.
The intrinsic nature is thus:
1) Mathematical scale invariance across the entire network.
2) Characteristics of hubs as being proteins which are highly conserved sequentially and functionally on examination because they have been evolutionarily constrained.
Scale-invariant biological networks can be easily explained by evolutionary theory as it stands today. Evolutionary theory also predicts that the proteins within the hubs show the least amount of average genetic drift as compared to non
-hub proteins.
THAT has been proven using, again, mathematics of sequence comparison.
So, let's go to the meat of the question, then.
Can ID explain why systems claimed to be IC are hubs? Or, in other terms, why can't non-hubs be IC?
It's not enough to answer that they're hubs because they're IC or vice versa. ID has to explain why IC protein groups cannot be found outside the stellate centers of scale-free networks. It is also not sufficient to say that IC proteins are in the center of stellate clusters because they have so many IC partners, because that's not true. Not all hub protein partners are necessary for that organisms' survival. Just some certain ones are. See the Barabasi paper linked in the previous post for details.
If ID did not require sequential evolution, then IC systems /should/ be found outside the hubs of scale-invariant networks whose associating proteins are many more than what IC needs to allow. You should be able to find little outlying clusters of IC systems in the network space who function by themselves on peripheries like little machines. We simply do not find those. They do not exist.
And the network, in ID, should NOT exhibit constraint. An ID system has no need to stay within the tight bounds of a scale-invariant network, yet we see these scale-invariant networks jealously maintaining their mathematical structure. I posit that the scale invariance, the mathematical quality of these networks, is proof positive that sequential and stepwise addition of components has occured because scale free biological networks will not allow any other mechanism of generation.
I'm asking if you can think of any other mechanism, outside of those thought to occur in evolution, which would go far to explain the mathematical quality of scale invariance in biological systems. The mathematical quality should be the context, meaning the analytical graph quality of the network map.
Regards,
Deanne
[ 19 September 2002, 23:26: Message edited by: Deanne M. Taylor ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 19. September 2002 23:05
Hi Andy,
You write: This smacks of a straw-man argument to me. Who are these people who "insist" that ALL (emphasis mine) knowledgeable, intelligent, sane, and honest people would admit that the immune system evolved by natural selection? The only possible candidate I could think of offhand would be Richard Dawkins.
Perhaps you are right and Dawkins' views are simply a freakish anomaly (my experience makes me think not, but hey, I'm jaded).
Okay, we can all thus agree that someone can be knowledgeable, intelligent, sane, and honest, yet also have real doubts that the immune system did in fact evolve by natural selection.
[ 19 September 2002, 23:12: Message edited by: Mike Gene ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 20. September 2002 09:22
I asked Charlie:
quote:
What evidence shows that the proto-RAG transposon would insert only into a primordial Ig locus, and be expressed only in the cell types where that locus was active?
He replied:
quote:
None - there is no such requirement. Transposons insert themselves pretty much anywhere. They are probably doing that, right now as I type, in my genome and yours. In most cases they do no harm, in others they cause disease, in a few cases, they may create new functional combinations by moving stuff around.
The examples you provided above were diseases (cancers). The "new functional combinations," however, appear to be hypothetical. That's my point. Looks as if I've got to take Inlay's word for it, and yours, that transposons landing randomly throughout the genome will do what is claimed, vis-a-vis adaptive evolution.
I asked Charlie:
quote:
2. What evidence shows that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon?
He replied:
quote:
None, but that's not a scientifically testable question, is it?
Of course it's testable. Transposons can be introduced into cells, and we observe what happens to those cells. Do they evolve "new functional combinations" or not?
So, on this second point, you've again provided no evidence for your claim.
I said:
quote:
The “vague, unsatisfiable requirements” aren’t mine. They are set by the theory of natural selection itself. In the thread above, Yersinia argued that selection explained the origin of the immune system. To support that claim, he needs to provide the relevant evidence.
He replied:
quote:
Wrong: the theory of natural selection does not require that every single trait be testable for its adaptive value. That's just an artificial hurdle set by skeptics of the theory.
I didn't say "every single trait" -- but for the trait in question, claimed to have evolved via natural selection, one certainly must show (a) variation, (b) a fitness difference, and (c) heritability. That's population genetics 101.
Skip the rest.
[ 20 September 2002, 09:25: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 20. September 2002 10:54
Yersinia wrote:
quote:
But: the favoring of diversity in immune system in molecules (and conversely, in the external molecules of diseases) is not a supposition, rather it is a trend we see just about everywhere we look -- the adaptive immune system is just one way to increase diversity. We can tell that positive selection for diversity of immune system genes has been operating for at least the last few millions of years based on substitution rates, and I bet that we'd find this wherever we look.
Your argument is question-begging, to wit: The vertebrate immune system generates greater molecular diversity; therefore greater molecular diversity was positively selected. Using that sort of reasoning, however, one could equally well argue that "simpler" immnue systems were also positively selected (they're extant) or no immune systems at all.
I asked:
quote:
My original comment/question (2) was poorly phrased. Let me try again. Presumably the proto-RAG transposon could land anywhere in the genome, and be expressed (or not). What evidence shows that the proto-RAG transposon would insert only into a primordial Ig locus, and be expressed only in the cell types where that locus was active?
Yersinia replied:
quote:
Sure, the transposon could insert anywhere, just like a mutation could happen anywhere. It does not have to happen miraculously just in that one specific spot.
But it does. That's Inlay's hypothesis. That is, the insertion must occur at least in the original Ig locus, and must provide a selectable function.
I asked:
quote:
2. What evidence shows that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon?
Charlie replied "none," and added that this was not a scientifically testable matter.
Yersinia, however, replied:
quote:
In the first place, all we're talking about is slight variations in the resulting receptor, and we know that cells make use of a diverse array of non-rearranging receptors. In the second place, as noted before these receptor genes are particularly prone to duplication and modification (redundancy is everywhere), all we're talking about here is taking one of these many copies and giving it a slightly different type of mutation. In the third place, if a few of the lymphocyte-like cell lineages had screwed up receptors initially, so what? (Particularly if this proto-lymphocyte was expected to have the "if my receptor finds something I will replicate" quality, rafe or charlie can comment on if cells with non-rearranging receptors tend to have this)
I'll let you two sort this out. Maybe Yersinia can educate Charlie, who I think is an immunologist.
I wrote, in response to the use of "selection pressure" as a cause:
quote:
Bacteria have nothing resembling the vertebrate immune system, and they're doing just fine.
Yersinia replied:
quote:
Bacteria are not large multicellular critters with blood streams and stomachs and such to infect.
Well, sure, but that's just the point. Your use of "selection pressure" as a cause is strictly post hoc. If vertebrates have complex immune systems, then selection favored the evolution of such. Organisms without such systems aren't under the same selection pressure -- and the evidence for this is their lack of the system.
I'll have to leave this thread until after 9/27, to prepare for my lectures at the ID conference being held at the University of San Francisco:
http://www.ideacenter.org/usf2002/
I'd encourage anyone in the Bay Area who wants to say hello to stop by!
[ 20 September 2002, 11:19: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
pz
Member
Member # 400
|
posted 20. September 2002 11:27
pz's post has been deleted for breaking several board rules. He/she managed to do this with a one sentence quote and a one sentence comment.
[ 20 September 2002, 11:38: Message edited by: Moderator ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 20. September 2002 11:35
quote:
The examples you provided above were diseases (cancers). The "new functional combinations," however, appear to be hypothetical. That's my point. Looks as if I've got to take Inlay's word for it, and yours, that transposons landing randomly throughout the genome will do what is claimed, vis-a-vis adaptive evolution.
Well, you guys decide: do you want to be "Pubmed-bombed", or are you going to do your own literature searches? In fact, the role of transposons in genome evolution is well known since Barbara McClintock's times, and I actually thought it was rather uncontroversial. Anyway, as far as new specific functional combinations arising from transposase activities, the literature is quite rich, you can start with:
- Flower color variation in the morning glory (reviewed here), including the transpositional insertion of unrelated exons into genes;
- alteration of gene expression by transposon-carried transcriptional regulatory elements in mammalian cells (reviewed here);
- changes in the coding sequences of Ascobolusspore color genes due to imprecise transposon excision (note the parallels with VDJ recombination junctional variability, see below).
quote:
Of course it's testable. Transposons can be introduced into cells, and we observe what happens to those cells. Do they evolve "new functional combinations" or not?
So, on this second point, you've again provided no evidence for your claim.
That's not what you asked at all. You asked to show "that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon." This is clearly untestable. For the generation of new functional coding sequences by transposon excision under experimental conditions, though, you can try to read the paper on Ascobolus cited above.
quote:
I didn't say "every single trait" -- but for the trait in question, claimed to have evolved via natural selection, one certainly must show (a) variation, (b) a fitness difference, and (c) heritability. That's population genetics 101.
Uh? Really, how can you expect to show anything like that for traits that formed in extinct organisms, living in unknown ecosystems, through stochastically unique, irreproducible mutations in ancestral genes we only know by their current descendants? That's not population genetics 101, it's elementary logic. [Incidentally, what nic was doing in his answer to this question was simply to show that such selective advantage is entirely reasonable - you first ask for untestable scenarios, then get upset because they are untestable].
The point that you, Dembski, Behe and many others repeatedly fail to appreciate is that one doesn't have to (indeed, one cannot, short of the invention of time travel) show a specific, quantifiable selective advantage for all the steps of any purported evolutionary pathway. Just like one doesn't have to put up a physical demonstration of the separation of continental land masses to prove that plate tectonics is a valid scientific explanation for the origin of current continents, or travel to Jupiter to show that the laws of physics are the same there as here.
That's tough for a philosopher to accept, I guess. Maybe you'll never be satisfied that VDJ recombination arose by standard naturalistic evolutionary means, or that stars evolve from slowly coalescing gas clouds instead of active squeezing by divine hands. You are asking from science an omniscence it cannot give, instead of the intellectually satisfying specific explanations, testable within current experimental limitations, it can (and does) provide.
[ 20 September 2002, 11:51: Message edited by: charlie d. ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 20. September 2002 13:16
I had intended to leave this thread temporarily, but...
Charlie wrote:
quote:
as far as new specific functional combinations arising from transposase activities, the literature is quite rich, you can start with:
- Flower color variation in the morning glory (reviewed here), including the transpositional insertion of unrelated exons into genes;
- alteration of gene expression by transposon-carried transcriptional regulatory elements in mammalian cells (reviewed here);
- changes in the coding sequences of Ascobolusspore color genes due to imprecise transposon excision (note the parallels with VDJ recombination junctional variability, see below).
PubMed literature bombing at its best. Let's revisit the original hypothesis on the table, namely, Matt Inlay's:
quote:
A transposon containing the RAG genes and flanked by two RSSs integrates itself into the gene for a primordial antigen-receptor gene, splitting it into two gene segments (V and J). The locus itself is transciptionally inactive in most cell types and prevents the expression of the RAG genes and removal of the integrated transposon. However, in a lymphocyte-like cell, the locus becomes transcriptionally active, and the RAG genes express themselves and remove the transposon, reuniting the two gene segments. The imprecise joining process generates a level of receptor diversity that favors the organism bearing this transposon, and its descendants thrive with the increased immune capacities.
Now, which of the citations above, Charlie, supports this scenario? The "new functional combination" at issue is not morning glory color, or spore color in a fungus: it's immune function. Do you have any evidence showing that transposon insertion into immune system loci is anything other than neutral or deleterious?
“Literature bombing” is the practice of appearing to answer a question by misdirection. I’ll conjecture that the evidence indicates that mutations (including transposon-induced mutations) to immune system loci are either neutral or deleterious, and that “new functional combinations” have not been observed.
Which again leaves one having to take Inlay's word for it.
Charlie wrote:
quote:
You asked to show "that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon." This is clearly untestable.
But Matt Inlay claims to know that the mutations are not only possible, but actually occurred and were spread in the population via natural selection. If this claim is untestable, however, why should one accept it as true?
I wrote:
quote:
I didn't say "every single trait" -- but for the trait in question, claimed to have evolved via natural selection, one certainly must show (a) variation, (b) a fitness difference, and (c) heritability. That's population genetics 101.
Charlie replied:
quote:
Really, how can you expect to show anything like that for traits that formed in extinct organisms, living in unknown ecosystems, through stochastically unique, irreproducible mutations in ancestral genes we only know by their current descendants?
Then how do you know that natural selection was the causal process at work?
Skip the rest as irrelevant badinage. Now I will quit the thread (for the time being, anyway). Last word to Charlie or whoever.
[ 20 September 2002, 13:27: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 20. September 2002 14:55
Paul:
I am not sure I understand you correctly, but if I do, I truly am surprised at the depth of your misunderstanding of how science works. By your token, science should be made only of limited direct observations, with no place for inference. By your refusal to accept any concept that is not directly testable or observable, you would basically throw away large parts of modern science: most of astronomy, cosmology, paleontology, and geology, quite a bit of particle physics and biology, heck even archeology and history are entirely untrustworthy for you.
As for the specifics, you originally claimed that the ability of transposons to generate new functional genetic properties was entirely "hypothetical". I simply showed you it is not, and that changes in gene structure, sequence and regulation by transposons have been shown in a variety of systems (call it "Pubmed bombing", if you like, I think it's supporting one's claims with evidence). Scientifically, it is irrelevant if the evidence comes from flower color or fungus spore genes, just like it is irrelevant if evidence for gravitation's role in determining Jupiter's moons orbits comes from studying falling apples and cannon balls.
The same misunderstanding shows up here:
quote:
But Matt Inlay claims to know that the mutations are not only possible, but actually occurred and were spread in the population via natural selection. If this claim is untestable, however, why should one accept it as true?
Because we know transposon-mediated mutations like those proposed by Inlay do happen, and that variability in transposon excision junctions of an antigen receptor gene (as shown by the Ascobolus experiments) would have generated diversity in the sequence of the antigen receptor genes, and that diversity in the sequence of antigen receptor genes, at face value, seems to be a good thing to have, since many organisms go to great lengths to get as much antigen receptor variability as they can (indeed, often, much more than they can use), and indeed can't seem to do very well without. Finally, because, while there is plenty of evidence of the role or RM/NS in evolutionary mechanisms, there is no evidence for non-naturalistic mechanisms, especially for ID. As of now, the model described by Inlay is the best model available for the origin of antigen receptor gene rearrangements, it's entirely consistent with the known facts, allows for testable predictions (some of which were tested and confirmed), and does not make recourse to unprecedented unexplicable mechanisms. You don't have to accept it as true, but if you have a better one, bring it forth. Until that happens, or evidence shows the model wrong, it will stand.
Maybe it sounds like "badinage" to you, but that's science.
[ 20 September 2002, 15:15: Message edited by: charlie d. ]
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 20. September 2002 15:50
Paul Nelson has made several more demonstrably wrong assertions about the science which I will attempt to get to soon.
However, one point of confusion in the discussion that Nelson has pointed out is that Charlie D. and I appear to be disagreeing on the testability of the transposon hypothesis, as Nelson points out when quoting Charlie D:
quote:
[Charlie D]
quote:
You asked to show "that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon." This is clearly untestable.
But Matt Inlay claims to know that the mutations are not only possible, but actually occurred and were spread in the population via natural selection. If this claim is untestable, however, why should one accept it as true?
I, on the other hand, would say that something like "that the normal function of the primordial Ig locus -- i.e., its function prior to the RAG insertion -- would not be disrupted or compromised by the RAG transposon" is testable.
I think that Charlie D and I are using the term "testability" in different ways here. Perhaps, Charlie D., you meant something like "we can't perform a direct in-the-lab experimental replication of the model, because the event occurred 500 million years ago in a long-extinct organism". I would agree with this, but I think that in science, model "testability" often has the more general meaning of "the likelihood of the model can be assessed by seeing if predictions of the model are supported by observations."
I think that the transposon hypothesis is clearly testable using this latter definition. And furthermore, Nelson appears to agree as he has made objections which are similarly testable and even made some statements that I think are refutable based on current evidence.
(Checking the thread, Charlie D., you appear to have just done this on several points, i.e. pointing out observations in extant observations which raise the likelihood of the transposon model being correct, and decrease the likelihood of Paul's counterarguments.)
Would you agree that the "in the lab replication of the scenario" testability is distinguishable from this more general form of testability?
yersinia
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 20. September 2002 16:23
Of course scientific inferences can (indeed, have to) be testable, if not directly, indirectly by the predictions they make. What I was referring to as untestable, however, was this request from Paul:
quote:
I didn't say "every single trait" -- but for the trait in question, claimed to have evolved via natural selection, one certainly must show (a) variation, (b) a fitness difference, and (c) heritability. That's population genetics 101.
Obviously, for the trait in question, i.e. a newly transposon-disrupted antigen receptor precursor gene in an extinct protovertebrate, these expectations are just as absurd as claiming that to sufficiently prove plate tectonics' role in forming continents we need to witness the formation of a new continent.
I think this is not just a problem of communication, but a real flaw in Paul's understanding of science. For instance, he later dismisses evidence of the role of transposons in generating novel genetic functions by calling it "misdirection", while one of the papers I quote specifically shows that a transposon excision causes the same type of junctional diversity that is predicted by the transposon-excision model of antigen receptor recombination, i.e. insertion and deletion of short nucleotide stretches, at detectable frequencies, causing changes in coding sequences. To me, as a scientist, this counts as a tested and confirmed prediction of the model. For Paul, it's simply irrelevant, because the "issue is not morning glory color, or spore color in a fungus: it's immune function."
So, your (and my, and every other scientist's) kind of "testability" of models is never going to satisfy Paul's standard of scientific "demonstration". For him, either we witness an evolutionary event (or faithfully reproduce it later), or we can't say it happened.
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 20. September 2002 17:39
Just how far Nelson is from even understanding what he is criticizing is demonstrated by these two comments where he *repeats* the errors I pointed out in a previous post without even bothering to disagree with my critique:
quote:
Your argument is question-begging, to wit: The vertebrate immune system generates greater molecular diversity; therefore greater molecular diversity was positively selected. Using that sort of reasoning, however, one could equally well argue that "simpler" immune systems were also positively selected (they're extant) or no immune systems at all.
Let me make it very simple: the evidence that diversity in receptor molecules is favored is not based on the observation that vertebrates have a mechanism to increase said diversity.
Off the top of my inexpert head, the evidence for the proposition "immune system receptor diversity is favored by natural selection" comes in a number of forms. Here is one:
Looking at gene families and duplications of Ig genes *and also* non-rearranging receptor genes in various organisms, and observing evidence of positive divergent selection on duplicated copies of genes by looking at synonymous vs. non-synonymous subsitution rates in these genes compared to those of other gene duplications/families
I have vague memories of studies of breeding preferences, immune system depression in inbred, reduced-receptor-diversity critters and such things which also bear on the question, perhaps someone else has this fresher in their minds however. I suspect that there are lines of evidence on this topic not dreamed of in our philosophies...
And, there is also support for the diversity-is-favored idea from theoretical analyses of probability-of-detecting-invaders-in-time. That is what this article is about for instance:
quote:
Cohn M. Immunol Rev 2002 Jul;185(1):24-38
The immune system: a weapon of mass destruction invented by evolution to even the odds during the war of the DNAs.
To avoid literature-bombing, here is just one article on selection-for-diversity topic:
quote:
Ota T, Nei M. Mol Biol Evol 1994 May;11(3):469-82
Divergent evolution and evolution by the birth-and-death process in the immunoglobulin VH gene family.
Immunoglobulin diversity is generated primarily by the heavy- and light-chain variable-region gene families. To understand the pattern of long-term evolution of the heavy-chain variable-region (VH) gene family, which is composed of a large number of member genes, the evolutionary relationships of representative VH genes from diverse organisms of vertebrates were studied by constructing a phylogenetic tree. This tree indicates that the vertebrate VH genes can be classified into group A, B, C, D, and E genes. All VH genes from cartilaginous fishes such as sharks and skates from a monophyletic group and belong to group E, whereas group D consists of bony-fish VH genes. By contrast, group C includes not only some fish genes but also amphibian, reptile, bird, and mammalian genes. Group A and B genes were composed of the genes from mammals and amphibians. The phylogenetic analysis also suggests that mammalian VH genes are classified into three clusters--i.e., mammalian clans I, II, and III-and that these clans have coexisted in the genome for >400 Myr. To study the short-term evolution of VH genes, the phylogenetic analysis of human group A (clan I) and C (clan III) genes was also conducted. The results obtained show that VH pseudogenes have evolved much faster than functional genes and that they have branched off from various functional VH genes. There is little indication that the VH gene family has been subject to concerted evolution that homogenizes member genes. These observations indicate that the VH genes are subject to divergent evolution due to diversifying selection and evolution by the birth-and-death process caused by gene duplication and dysfunctioning mutation. Thus, the evolutionary pattern of this monofunctional multigene family is quite different from that of such gene families as the ribosomal RNA and histone gene families.
On the second issue:
quote:
I asked:
quote:
My original comment/question (2) was poorly phrased. Let me try again. Presumably the proto-RAG transposon could land anywhere in the genome, and be expressed (or not). What evidence shows that the proto-RAG transposon would insert only into a primordial Ig locus, and be expressed only in the cell types where that locus was active?
Yersinia replied:
quote:
Sure, the transposon could insert anywhere, just like a mutation could happen anywhere. It does not have to happen miraculously just in that one specific spot.
But it does. That's Inlay's hypothesis. That is, the insertion must occur at least in the original Ig locus, and must provide a selectable function.
Incorrect! You have been assuming that transposon insertion occurred just once in just the right spot. That's not Inlay's hypothesis. Inlay's hypothesis is that in a population of organisms (proto-jawed fish), transposon insertions (and the other kinds of mutation) were occasionally happening at random spots in various genomes. Transposons likely got inserted all kinds of places the genomes of various individuals. Sometimes this would be a detrimental mutation and that egg would not develop, or that organism would have some other defect and get selected out, and therefore these would be eliminated from the popultion. Often times transposons would get inserted into noncoding DNA and these would be more-or-less neutral. All that Inlay's model proposes is that among all of these random insertion events, very occasionally a transposon inserted into (one of the many) non-rearranging receptor genes that this organism had (each organism in the population would have had many such receptors, as (IIRC) all bilaterans examined thus far have large numbers of non-rearranging receptors) and converted into a slightly-variable receptor that conferred a slight reproductive advantage. Sooner or later, one of these transposon insertions would succeed in spreading to fixation in the population, and this would be the ancestor of the RAGs that we see today.
...in other words, like any other mutation, a very large number of "misses" makes up for the small proportion of "hits". We do not have to suppose that harmful mutations never occur or that the model's transposon insertion was the only one that ever occurred.
yersinia
[ 20 September 2002, 17:45: Message edited by: yersinia ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 20. September 2002 19:25
Nic: Just how far Nelson is from even understanding what he is criticizing is demonstrated by these two comments...
Don't you think you can make your points without trying to make someone else's "understanding" an issue?
IP: Logged
|
|
rafe gutman
Member
Member # 134
|
posted 21. September 2002 00:37
miscellaneous comments by paul (all boldface mine):
quote:
In his talk.design article, Matt Inlay said that the descendants of the original RAG mutant "thrive with the increased immune capabilities." But how does he know this?
then later
Looks as if I've got to take Inlay's word for it, and yours, that transposons landing randomly throughout the genome will do what is claimed, vis-a-vis adaptive evolution
then even later
I’ll conjecture that the evidence indicates that mutations (including transposon-induced mutations) to immune system loci are either neutral or deleterious, and that “new functional combinations” have not been observed.
Which again leaves one having to take Inlay's word for it.
and finally
But Matt Inlay claims to know that the mutations are not only possible, but actually occurred and were spread in the population via natural selection. If this claim is untestable, however, why should one accept it as true?
paul, i think you misunderstand the point of a model. a model is not meant to be the finished product. every feature of the model need not be proven beyond a reasonable doubt. it is not the final answer. the point of a model is to explain the existing data and tie it all together into one cohesive unit. you keep insisting that the
author claims his model as the truth, as what actually happened. this is not what a model is. besides, at the beginning of the paragraph of the model you quote, he writes: "A more detailed model could proceed as follows..." why are you stuffing words into people's mouths?
as to your complaints, there will of course be gaps in our knowledge, and a good model will try to explain those gaps with testable hypotheses. you can call that "storytelling" if you like, but this type of storytelling can lead to actual research. if nothing else, it gives us a direction to go in. again, i have to point out that ID has contributed nothing to this discussion.
you're not forced to believe the model as truth. a model like that should be viewed skeptically, or rather, tentatively. one of the key pieces of evidence for that model is still missing, and may never be found. however, until someone comes up with a better model, immunologists will consider it the best one available. right now, i don't think there is even a contender (even among evolution supporters).
now on to the details:
quote:
a portion of the model from the inlay article
A transposon containing the RAG genes and flanked by two RSSs integrates itself into the gene for a primordial antigen-receptor gene, splitting it into two gene segments (V and J). The locus itself is transciptionally inactive in most cell types and prevents the expression of the RAG genes and removal of the integrated transposon. However, in a lymphocyte-like cell, the locus becomes transcriptionally active, and the RAG genes express themselves and remove the transposon, reuniting the two gene segments. The imprecise joining process generates a level of receptor diversity that favors the organism bearing this transposon, and its descendants thrive with the increased immune capacities.
here's a run down of the general details of this model
1. the RAG genes and RSSs were originally a transposon, and the RAG proteins had transposase activity.
2. a gene existed in the ancestor to the jawed vertebrates that encoded a receptor with at least one immunoglobulin domain
3. this gene was only expressed in a somatic lineage, like hemocytes
4. the receptor could activate a pathway somehow involved in immune-related function
5. the RAG transposon integrated into the receptor near the binding site-coding region.
6. once inserted, the transposase
would be under the receptor's transcriptional control.
7. the RAG's ability to reassemble the receptor gene once activated was sloppy, and made slight sequence changes upon reassembly.
8. those changes were beneficial to the host
and the evidence supporting it:
1. there seems to be ample evidence suggesting that the RAG genes were once transposases, do you dispute this? the fact that they have transposase activity (in vitro) basically says it all.
2. the immunoglobulin domain is well-represented in nearly all organisms (even prokaryotes, i think). because its structure is so stable, a large portion of cell-surface molecules possess it, even some related to immunity (outside of the jawed vertebrates). however, no direct homologue to the antibody proteins have been found. so here is a prediction of the model: a non-rearranging receptor with an immunoglobulin domain existed in the ancestor to the jawed vertebrates. of course, this ancestral gene may no longer exist, but if it is ever discovered, that would fill in a major hole.
3. a lot of genes have tissue specific expression. as for the extant antibody genes, they are heavily regulated by enhancers (because of the need to prevent premature recombination, or recombination on both alleles). there's no reason to think that the enhancer wasn't there in the ancestral receptor (such as, if it was located between the V and J segments).
4. see 2.
5. this is the only step that appeals to chance. as yersinia said, there could have been many integrations by transposons (our genome is littered with remnants of retrotransposons). all it takes is one to be beneficial, and it will become fixed into the population.
6. a common molecular technique is transgenics, where whole genes are inserted into the chromosome randomly. one of the main problems with this approach is the lack of consistency in transgene expression. this is because the transgene becomes subject to the local transcription control. unless the transgene has a really powerful promoter, it's uncertain whether or not it will be expressed correctly.
7. the current mechanism of RAG-mediated recombination is sloppy. one of the reason is through the generation of hairpin loops at the two exposed ends. this is intrinsic to the activity of the RAG and is common to transposases (in fact, this is one of the observations that led scientists into thinking the RAGs were originally transposons).
8. here the evidence may never be to your satisfaction, but it's very reasonable to suppose that a high rate of mutation would benefit immune receptors. why do you think new flu vaccines come out every season? because the influenza's receptors can mutate very rapidly to avoid host recognition. furthermore, they don't need to mutate drastically and change the total structure of their receptors, only just enough to hamper the host's receptor's binding ability. if during tthe lifetime of the first RAG-integrated organism, one of the thousands to millions of hemocytes alive at the time form a receptor beneficial to the host, this will increase the host's chances for survival and the nature will have something to select upon.
maybe this requires more evidence to convince you, but is it really that unreasonable?
[ 21 September 2002, 00:44: Message edited by: rafe gutman ]
IP: Logged
|
|
|
Topic Closed
Printer-friendly view of this topic
