|
Author
|
Topic: Evolving Inventions
|
charlie d.
Member
Member # 159
|
posted 31. January 2003 23:02
Nelson:
you are closer than you think. Indeed, the whole point of that article is that wing re-evolution may not be such a big deal after all. The question is, why?
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 31. January 2003 23:04
It's not re-evolution if they don't have re-build the genes that develop the wings, which is what re-evolution is. Evolving the same pathway different times would be re-evolution. This may be the simple turning off and on of the same genes. The information is still there. The question is how did they get the information in the first place...
[ 31. January 2003, 23:05: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 31. January 2003 23:06
You mean, re-winged stick insects would need to have an entire new set of complete genetic intructions for wing development (possibly different from the original ones)?
[EDIT: Oh, now I see. You think it's simply a matter of gene regulation. But what turns the wing-making genes on and off, if not other genes? And how are those regulated? See the point? You may be farther than I thought initially]
[ 31. January 2003, 23:09: Message edited by: charlie d. ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 01. February 2003 21:48
Charlie:
You mean, re-winged stick insects would need to have an entire new set of complete genetic intructions for wing development (possibly different from the original ones)?
Nelson:
The article thinks that this might be a violation of Dollo's law because the same pathway was accomplished multiple times. Insect researchers correctly view this as unlikely. Think about it. If evolution is essentially random then what we have here is clearly the random evolution of a complex system, it's loss, then the random evolution of the same complex system again? And again? What is most likely scenario, and most widely accepted, is the simple turning off/on of the same genes that control development of different morphological structures. However, this leaves the question of how the information got there in the first place.
Charlie:
[EDIT: Oh, now I see. You think it's simply a matter of gene regulation. But what turns the wing-making genes on and off, if not other genes? And how are those regulated? See the point? You may be farther than I thought initially]
Nelson:
No I'm sorry I don't see the point. Feel free to explain, making sure you include the names and functions of these other genes that turn the wing making genes on/off.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 01. February 2003 22:04
Nelson:
many, if not most genes in any developmental pathway (and probably the wing developmental pathway as well) are genes that turn other genes on or off (ie, they encode for transcription factors, or for signaling molecules that result in gene expression changes). So, saying that the pathway is turned on or off doesn't really add much: I am not a development guy, but off the top of my head I can think of a number of genes involved in wing development whose job is to turn other genes or or off: fringe, notch, wingless, frizzled, hedgehog (I know, drosophila people have all the nomenclature fun!)- and there are more I am sure.
What I do not understand from your hypothesis is: say the critical change is in a trancription factor gene somewhere, how do you propose the gene is turned off? A gene mutation, or some other mechanism (and which)?
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 01. February 2003 22:34
Charlie:
many, if not most genes in any developmental pathway (and probably the wing developmental pathway as well) are genes that turn other genes on or off (ie, they encode for transcription factors, or for signaling molecules that result in gene expression changes). So, saying that the pathway is turned on or off doesn't really add much: I am not a development guy, but off the top of my head I can think of a number of genes involved in wing development whose job is to turn other genes or or off: fringe, notch, wingless, frizzled, hedgehog (I know, drosophila people have all the nomenclature fun!)- and there are more I am sure.
Nelson:
Yes but I was hoping you would tie this in relevantly to whatever I was close to or far from.
Charlie:
What I do not understand from your hypothesis is: say the critical change is in a trancription factor gene somewhere, how do you propose the gene is turned off? A gene mutation, or some other mechanism (and which)?
Nelson:
Probably.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 02. February 2003 01:26
Nelson,
The pod-1 result I cited is a "personal communication" level result. I could possibly be remembering it wrong, too.
However, there's a published example that illustrates the same point: the cold-sensitive allele ye60 of the pod-2 gene (also involved in polarity) gives variable penetrance in its lethality as you upshift the temperature at different times during development. The reference for that one is Tagawa et al., Developmental Biology 233:412-424 (2001).
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 02. February 2003 02:05
Thanks Rex,
I'll probably pick up the paper this week and make some comments on it soon.
[ 02. February 2003, 02:07: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 02. February 2003 09:39
Cool, Nelson. Next step.
We agree that there is a whole genetic pathway for wing development, with probably dozens of component genes, and that a mutation in a (possibly, regulatory) gene somewhere along the pathway must have turned wing development off in a branch of the stick insects family. They live happily for million of years without wings.
What happens to the other genes in the pathway, while they are being turned off and useless? Do they just wait, intact, for future re-evolution, or would you expect them to eventually acquire crippling mutations (as we commonly see happen in pseudogenes, for instance)?
[ 02. February 2003, 09:42: Message edited by: charlie d. ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 02. February 2003 19:34
Charlie:
We agree that there is a whole genetic pathway for wing development, with probably dozens of component genes, and that a mutation in a (possibly, regulatory) gene somewhere along the pathway must have turned wing development off in a branch of the stick insects family. They live happily for million of years without wings.
Nelson:
This point was never in contention. Whats in contention is whether this constitutes a re-evolution of the entire pathway again, and again, and again, which is unlikely. Or if it's just that the instructions for building a wing got turned off in some lineages. Loss of structures isn't very impressive at the phenotypical level.
Charlie:
What happens to the other genes in the pathway, while they are being turned off and useless? Do they just wait, intact, for future re-evolution, or would you expect them to eventually acquire crippling mutations (as we commonly see happen in pseudogenes, for instance)?
Nelson:
Charlie, we're going around in circles here. I've already discussed this several times (and it has been discussed by the report itself).
quote:
Researchers assumed wings could not come back once lost as the genes needed to create them would mutate beyond repair once the wings disappeared. But Whiting says there is evidence from the fruit fly Drosophila that the same genes contain instructions for forming wings and legs.
If the same were true for stick insects, there would be an evolutionary pressure to stop wing genes from mutating, even in the insects that did not have wings. Those genes could then be turned back on in the future.
This is quite well understood and quite likely an explanation. Also, another report notes:
quote:
The authors note that the new study does not speak to the biggest question about such complex features - their origin. That is, no cycles of losing and regaining of wings can explain how insects first evolved their wings.
There is an answer to your question here. Earlier you stated:
quote:
So, John, in your scheme of how developmental innovations can (actually, cannot) arise, how would you interpret the multiple re-evolution of wings in stick insects, and what it is possibly telling us about the first appearance of wings in insects?
It tells us absolutely nothing about the first appearance.
This is because:
quote:
Researchers said the findings suggested that the genetic machinery remained intact but latent in stick insects, possibly because that machinery is involved in developing features that the wingless insects continue to grow, such as legs. The team noted that the same could be true of other genetic machinery for other complex features.
Developmental biologists are kind of shrugging this off. They see ectopic structures grow where they shouldn't grow all the time.
[ 02. February 2003, 19:48: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 02. February 2003 19:55
See, you are close.
Let's recapitulate where we stand now. In some lineages of stick insects, some genes in the wing development pathway are mutated and wings do not form anymore. Yet, most if not all the other genes in the pathway do not "drift" away into obsolescence by accumulating more and more mutations, but remain functional because they are involved in other functions (making legs, or whatever else). [Incidentally, this of course means that those specific genes in the wing development pathway are actually turned off only in wings (or, to be precise, in the cells that would normally give rise to the wing), while they are still active in other organs performing other functions.]
So, if you just think about it, this very possibly gives us some hint about how the wing development pathway came to be put together in the first place, in the original non-winged insects, right? How does that relate to the PCP pathway John was talking about earlier in the thread? How would you go about testing that possibility?
[ 02. February 2003, 19:57: Message edited by: charlie d. ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 02. February 2003 20:01
Some preliminary thoughts about pod-1. Just to review, James stated:
quote:
You can see how mutating certain key genes would simply destroy the developmental process...Since the developmental process consists of a complex system of interacting genes, small changes cannot re-wire the entire system or produce the system de-novo. This is why I argue that fundamentally, the developmental process is fixed.
Rex stated:
quote:
Yet if you inhibit the function of [pod-1] with RNAi, despite a reduction in function of the gene, worms can sometimes survive (albeit sometimes somewhat messed up).
I have not read the paper on the cold sensitive pod mutants, but I think that sensitivity to low tempertures is not as impressive as reducing the function of a gene and still having a viable organism. After all, we do have lethality with the cold-sensitive pod. I find it very difficult to believe that reduction of function of a polarity gene would still yield a viable organism. For example, the author of the paper cited above stated in a review:
quote:
"If the asymmetrical division is screwed up, the embryo is never going to make it," says Akiko Tagawa, a graduate student and the lead author of the paper, which also includes Chad A. Rappleye, another graduate student working in Aroian's laboratory. "It quickly dies."
The important thing to do at this time is carefully dissect the function of these developmental genes and see which is key, and what would destroy what. I also intend on reading James' paper, as I unfortunately havn't as of yet.
[ 02. February 2003, 20:59: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 02. February 2003 21:07
Charlie:
So, if you just think about it, this very possibly gives us some hint about how the wing development pathway came to be put together in the first place, in the original non-winged insects, right?
Nelson:
No it doesn't. As the reports I have been quoting specifically suggest, the information for wing development was and is still there. So we still have no idea how we get wing-information from non-wing information. All that we see here is the turning off and on of wing information. This is why the reports specifically stated that this is irrelevant to the origin of wings.
Charlie:
How does that relate to the PCP pathway John was talking about earlier in the thread? How would you go about testing that possibility?
Nelson:
Ironically, both are related, in that neither are the multi-part examples of Darwinian inventiveness that Dr. Bracht was referring to.
[ 02. February 2003, 21:13: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 02. February 2003 21:52
Nelson:
we have just agreed that most of the genes involved in wing development are very likely involved in other functions as well. For this reason, even in insects that have lost wings due to mutation of some single gene along the pathway, the rest of the pathway doesn't quickly degenerate, but rather it maintains function and coherence, so that reactivation of the mutant gene (or its functional substitution with a new one) can lead to the apparently surprising phenomenon of wing "re-evolution".
In other words, the wing developmental pathway is made up in very large parts of elements with other functions. If this does not suggest to you the possible scenario that the original wing appearance was the result of co-option of disparate pre-existing developmental pathways with non-wing related functions, assembled and connected in new ways, with the addition of at best a few truly novel and independent wing-specific components, I am not sure what would.
Just like the PCP pathway, that was discounted by John as "a single part that got added to a pre-existing pathway to impart novel functionality", so wings may be the result of one or at best a few novel wing-specific factors being added to a vast network of pre-existing developmental pathways. Thus, wing appearance, though certainly representing a major evolutionary novelty, may not have been after all a very long walk away in the "hyperspace of possibility" (in John's words) for the original non-flying insect.
Luckily, an evolutionary scenario suggested by specific evidence such as this one (unlike the infinite variety of equally acceptable alternative design scenarios, from painstaking component-by-component tinkering to an all-in-one-shot "kazaam" event) also has the great advantage of being amenable to empirical testing, by establishing the identity and relationship of the components in the wing developmental pathway, their pleiotropic functions in other pathways, and the exact site of the inactivating and re-activating mutations in the various stick insect lineages.
It's going to be a fun ride. Can't wait.
IP: Logged
|
|
RBH
Member
Member # 380
|
posted 02. February 2003 22:53
John,
I apologize for coming to this thread late. I'd like to make one or two comments on your paper referenced in the ARN thread to which you provided a link in an early post in this thread. I did not wade through the whole ARN thread (in which I did not participate), so I hope you'll forgive me if some of what I say might repeat something there.
Genetic Algorithms
I am mostly interested, as I'm sure you know, in your comments on genetic algorithms. You say quote:
The encoding of the [GA] program completely determines the reference class in which that program may operate. The fixed parameters of the genetic algorithm completely determine the possibilities that may be generated and tested by the program. In the case of the Biomorph program, the number of genes is fixed and cannot be adjusted by the evolutionary process the program simulates.
This observation suggests that we may consider any genetic algorithm to be operating within a certain n-dimensional hypervolume, and certain fixed parameters completely determine that hypervolume ahead of time. Furthermore, any particular n-dimensional hypervolume is completely isolated and separate from any other m-dimensional hypervolume (m .n). As an obvious example, consider how three-dimensional objects are in a completely separate class from two-dimensional objects. Altering the width, length, or other characteristics of a circle will never convert the circle into a sphere; circles and spheres are from completely separate hypervolumes.
Those remarks reflect the same understanding of genetic algorithms that is embodied in MESA, the GA that you, Micah S. and William D. wrote to exemplify GAs to the ISCID audience. It is also consistent with a misinterpretation of Dawkns' "METHINKS IT IS A WEASEL" and his BioMorph program as somehow embodying the core capabilities (and limitations) of genetic algorithms in general. They don't. As I have argued elsewhere MESA is not a good representation of genetic algorithms as they are used in applied contexts, and it is an even poorer representation of the kinds of GAs that are used in simulations of biological processes.
As it happens, one can write (and people have written) GAs in which the kinds of parameters you refer to are not fixed. For example, AVIDA which was designed specifically to serve as a platform for artificial life research, allows its genomes to vary considerably in length, with insertion mutations to lengthen genomes and deletion mutations to shorten them. AVIDA's 'creatures' are self-reproducing assembly language programs. What is "fixed" in AVIDA is the assembly language, and that (like the physics and chemistry of DNA) constrains what is possible. However, the language used in AVIDA is Turing complete, so that is not a severe constraint.
Your description of Dave Thomas's presentation and your correspondence with him illustrates the same confusion that is embodied in MESA, a confusion between (1) GAs as search algorithms, and (b) GAs as models for (platforms for research in) biology. In the former case, it is obvious that a fitness function is supplied by the writer - after all, that's what it's being used for, to find optima on the fitness landscapes induced by the fitness function and the several evolutionary operators. So yes, in that kind of application, one defines a fitness function and uses the GA as a search engine.
In using GAs as platforms for biological research - evaluating conjectures and hypotheses, generating approaches to test in wetware, and so on - the constraints are much looser, and indeed some are absent. For example, one can study populations evolving in the absence of any extrinsic imposed fitness function in AVIDA.
This will sound like heresy to some, but biological evolution is not a search process. As conceived by evolutionary theory, it is a process in which variation and selection (along with the other evolutionary operators) in co-evolving populations (with the attendant dynamically deforming fitness landscapes) produce the appearance of search for optima, but that appearance is deceptive. To be sure, populations find sufficiently high local optima (or perish), but they are not searching for those optima. They find optima as a by-product of the operation of evolutionary operators, not as a goal. Hence analyzing biological evolution as a search problem leads to false analogies with human uses of search algorithms on both sides of the debate.
Finding Novelty - Resolving Technical Contradictions
In spite of the fuzz associated with the difference between "routine" problems and "inventive" problems, there are indications now in the GA literature that GAs can indeed find "inventive" solutions to problems. In particular, the hardware evolution literature is providing examples of novel and inventive solutions to problems that in some cases cannot be understood by the program's designers. For example, Layzell describes the evolution of an oscillator that (apparently) takes advantage of parasitic capacitance in wiring as the source of nonlinearity that is necessary for an oscillator to run. In other cases, the experimenters have no idea how a GA-designed circuit is performing its function. To be sure, in these instances there was some experimenter control: The experimenters supplied the physical components (transistors and switches) and defined the fitness function, but the GA invented solutions that were not only unlike those a human designer would come to, but were difficult or impossible for human designers to even understand!
If one is going to use evolutionary algorithms in general or genetic algorithms in particular to try to argue that biological evolutionary processes cannot generate novelty, produce inventive solutions, or create "specified information," one really needs to get out of the MESA/WEASEL/BioMorph mindset with respect to GAs. They are toys, primitive, limited, and unrepresentative of real GAs that are operating out in the real worlds of applications and of research. One must be aware of the kinds of counter-examples I've mentioned here. At the least they make the argument harder to sustain; in my view they make it impossible to sustain. The argument from analogy is tough enough to sustain without mis-reading the source of the analogy.
RBH
[ 02. February 2003, 23:07: Message edited by: RBH ]
IP: Logged
|
|
|
Printer-friendly view of this topic