|
Author
|
Topic: Ontogenetic Depth and the Origin of Animals
|
Frances
Member
Member # 169
|
posted 11. February 2003 12:02
Paul
You quote Weiss to say that quote:
The assumption of a point source is important to the entire theory, so we insist upon spontaneous generation in the beginning, but must forbid it at any other time. (pp. 199-200)
But perhaps you should place his statement in its complete context.
First of all Weiss reports that
quote:
The grandeur of Darwin’s view was “of life having been breathed by the Creator into a few forms or into one.”
So the Darwinian point source has become point sources. Furthermore Weiss points out that horizontal gene transfer has muddled the roots of the tree.
When you state that Weiss argues that "common descent is inviolate" and reference page 203 we find that it reads
quote:
What beyond shared ancestry is inviolate?
Forthcoming columns will look at where some of these issues stand at the verge of a new century, to see where they still seem sound, and where they
may need changing, if we wish to understand
evolution more clearly.
If Paul believes that common descent is not the best explanation for the data then by all means he should present his best case for an alternative interpretation of the data.
Paul then comments
quote:
But your use of other echinoderm taxa to provide "intermediates" between H. erythogramma and H. tuberculata -- which is exactly how Frances understood your post, BTW -- is such a hypothesis, and an illegitimate one.
but what did I say
quote:
It seems that there may be evidence for a transitional sequence after all? Paul seems to object though because these intermediates are not for Heliocidaris. This is surprising because it seems that Paul would accept an evolutionary pathway for other species which show similar retarded development but would hold heliocidaris to be a special case? Why? After all the evidence shows that such intermediates are possible for other species. Is Heliocidaris special in some form? Certainly the evidence shows that contrary to Paul's suggestion in his talk, developmental networks can be changed to lead to significant morphological variations. Which is the argument in the paper by Smith
Paul, you have yet to show that this is in illegitimate 'hypothesis'. What Yersinia has shown is that intermediates exist and how changes in timing of various events may be a likely mechanism to explain these. As I stated before
quote:
A good review article on the timing of embryonic development is "TEMPORAL PATTERN FORMATION
BY HETEROCHRONIC GENES" by Frank Slack and Gary Ruvkun. They show that heterochronic variation is wide spread and suggests that variation in the gene loci which determine the developmental timing may play important roles in evolutionary change. They study heterochronic mutations in C. elegans and show that the control of the timing is analogous to Drosophila. Neoteny is another interesting concept in which the retarded expression of (previously) juvenile characteristics in the adult stage can lead to morphological changes.
Thus when you state that quote:
The "viable intermediates" you mention take on that status only when one strings together unrelated groups, which is biologically nonsensical.
you seem to ignore that these 'intermediates' show how variation in timing events in development can lead to significant morphological variations.
But in the end it would be helpful if Paul were to present his own ideas on alternatives to Common Descent, an idea which is strongly supported by the evidence.
Finally I would like to point to another reference in which Paul remarks on Weiss
quote:
And one could find this same axiomatic interpretation of CD defended by (for instance) Keith Stewart Thomson (now at Oxford, then at Yale), or, more recently, by Kenneth Weiss of Penn State (see his article, "We hold these truths to be self-evident," in a recent issue of Evolutionary Anthropology). Weiss argues that evolutionary theory assumes what he calls “a point source” for all of terrestrial life, as a genuine a priori standpoint. This assumption is simply not up for grabs (i.e., testing).
I find the suggestion that 'this assumption is not up for grabs (i.e. testing)' is a non sequitur. The constancy of the speed of light may bean a priori standpoint but can be validated.
[ 11. February 2003, 12:16: Message edited by: Frances ]
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 11. February 2003 12:51
Paul writes:
"Then you (Argon) should have no trouble saying how you would test common descent, i.e., the theory that all organisms on Earth share a common ancestor. If you don't think that theory is true, however, explain why. Give the test that the theory has failed, or provide the evidence that it cannot explain."
Well, now this is starting to sound like Walter ReMine when he talks about "cornucopias". I would ask: In your discussions with Mike Gene and Michael Behe, how did they respond? Why are they still "evolutionists?" Why does Denton, whose book inspired many of the leaders of ID science, say that organisms are clearly related by a common chain of descent?
I don't know that *all* organisms on Earth share a common ancestor. I think there is too much haziness at what would be the root of such a tree to make that a simple determination. They certainly do share a very common biochemistry, but whether that's the result of massive horizontal transfer from disparate sources very early in the process (via "intelligent" or "unintelligent" mechanisms), or common descent from a single ancestor is hard to tell.
But let's give common descent a test at levels "above the hazy root." After all, you aren't just questioning common descent at the root of life but at the more recent branches as well. Take Lake Victoria cichlids as a collection of species that are thought to have a fish (and metazoan) common ancestry. Do any of them lack mitochondria (or mitochondrial sequences), use something other than the Kreb's cycle and have something that looks unrelated to hemoglobin for oxygen transport? Do they lack nucleoside triphosphates as the primary currencies in metabolism? If we sequence their mRNAs (assuming the species have mRNA), do we competely fail at predicting the peptide sequence of the proteins they encode using "common" coding tables? Are their amino acids mostly of the D-form? If we can answer "yes" to any of these questions, we would have the beginnings of a pretty good case that any "cichlid" with these traits probably didn't come from other cichlids (or eukarotes, for that matter). If aliens landed on this planet, how would you go about determining whether we shared a common ancestor? This is not a difficult exercise. Yersinia has provided additional tests and the reasons behind them.
Or test one's skepticism on dogs. How do we test their proposed common ancestry with wolves? So what if they can interbreed, is it beyond the power of a design to separate create subspecies that can interbreed? One can deploy radical skepticism at any time, but when is it reasonable?
And what of horizontal transfer? How do we identify horizontal transfer events and how do we recognize that some parts of a genome do not appear to have arisen via common descent? Howard Ochman's group has a nice paper about the estimated amount of unrelated DNA that has been sampled and incorporated by E. coli (~1.6 megabases see here). Of course, one can't identify the sources of all these sequences, so I suppose one could argue for the possibility that these genes may have spontaneously introduced in design events rather than acquired from other organisms. But either way, it appears that some tests for common descent within particular groups are quite viable and powerful. And common descent has failed in some instances. I realize that Paul is asking whether "generative entrenchment" (GE) has a similar power, but given that we have strong, auxillary evidence suggesting otherwise, it may be prudent to continue attempts at validating that criterion.
So to review Paul's questions:
1) It is indeed possible to refute common descent explanations. That which has been applied to genomic fragments can be applied to whole genomes and organisms as well. Switching the focus to whether common descent can be evaluated for all of life or at its deepest roots oes not get around the fact that the topic was a specific case of the evolution and relatedness of congeneric sea urchins. Beyond "GE", Paul has not countered the systematics data or the fossil record.
2) Paul has not described how GE would prevent a designer using modification with descent as a mechanism.
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 12:55
Yersinia, your last post doesn't answer my question. Let me try again.
I don't want to know when (under what evidential conditions) you would say that two or more groups share a common ancestor. I want to know when you would say that they don't share a common ancestor, and why.
Previously, you provided two possible (albeit counterfactual) lines of evidence that would persuade you H. erythrogramma and H. tuberculata were unrelated:
quote:
Well, if the sequences of the two species failed to group together against other randomly chosen species (e.g., one species groups with humans, another with liverworts), that would be a pretty good indication.
If echinoderm junk DNA encoded "made by Rael, 5 million B.C.", that too would be a good indication.
OK -- why? What is it about these patterns that, if we observed them, would lead us to infer discontinuity?
[ 11. February 2003, 13:36: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 13:02
Frances wrote:
quote:
you seem to ignore that these 'intermediates' show how variation in timing events in development can lead to significant morphological variations.
The hypothesis of "variation in timing" is a post hoc rationalization, and assumes common descent, the very point at issue. Try to keep distinct what you've assumed versus observed.
Kenneth Weiss is only an e-mail away. Ask him if he thinks common descent is, or is not, properly an axiom of historical biology.
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 13:10
Argon,
So common descent (as ordinarily understood) may, or may not, be true. Okay. I assume that unlike Futuyma, Gould, Dawkins, and others, you wouldn't call the theory a fact, whose truth it would be perverse to doubt. That's progress.
You asked:
quote:
In your discussions with Mike Gene and Michael Behe, how did they respond? Why are they still "evolutionists?"
Mike Behe has said to me that he holds common descent provisionally, but that if the theory turned out to be false, he wouldn't care. I've never discussed the matter with Mike Gene (we correspond about other things).
quote:
Why does Denton, whose book inspired many of the leaders of ID science, say that organisms are clearly related by a common chain of descent?
The last time I talked at any length with Denton, in China in June 1999, he spent a lot of time talking about what he saw as imperfections (e.g., the recurrent laryngeal nerve) and how these features were hard to reconcile with an optimizing designer. Mike's main reason for accepting common descent, however, is philosophical. As he sees it, discontinuity requires an intervening designer, and he doesn't like that.
[ 11. February 2003, 13:37: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 11. February 2003 13:17
Hi Paul,
I just laid out the prediction based on the hypothesized process of descent with modification. If it wasn't obvious before, if instead of a nested hierarchy we got some other pattern predicted by some other hypothesized process, then that other process would be supported by the evidence and common descent would not be. You could, for instance, imagine a process whereby sequence similarity correlated with habitat.
I suppose I may have forgotten to add that the real key is that different datasets should yield similar trees if common descent is responsible. E.g. the first 50% of each of our simulated sequences should produce the same tree as that derived from the second 50% of those sequences. Lineal common descent of two species would be falsified if independent datasets (e.g., two different genes) never gave similar trees -- e.g., gene1 groups sea urchin species A with humans and B with liverworts, gene2 with frogs and slime molds, respectively, etc.
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 11. February 2003 13:22
quote:
So common descent (as ordinarily understood) may, or may not, be true. Okay. I assume that unlike Futuyma, Gould, Dawkins, and others, then, that you wouldn't call the theory a fact, whose truth it would be perverse to doubt. That's progress.
The question is not whether it is perverse to doubt common descent at all -- depending on definitions, at extremely remote time scales this is a reasonable thing to doubt. Certainly lineal common descent is not universally true.
The question is whether or not it is perverse to doubt the common descent of two congeneric species of sea urchins which exhibit close sequence similarity on several measures.
[edit to add: and which fall together in a long standing nested hierarchy of echinoderms, deuterostomes, metazoans, eukaryotes, etc.]
[ 11. February 2003, 13:31: Message edited by: yersinia ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 13:42
Yersinia wrote:
quote:
Lineal common descent of two species would be falsified if independent datasets (e.g., two different genes) never gave similar trees -- e.g., gene1 groups sea urchin species A with humans and B with liverworts, gene2 with frogs and slime molds, respectively, etc.
Of course, but we'd also say why common descent would be falsified. That's the missing piece, which I'm hoping you'll provide (so that we can get back to the details of Heliocidaris).
What is it about the pattern you describe that would be impossible to reconcile with common descent?
Darwin knew the answer, incidentally.
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 13:50
Argon wrote:
quote:
Paul has not described how GE would prevent a designer using modification with descent as a mechanism.
Paul has not described it, because absolutely nothing (including GE) prevents a designer from doing whatever he pleases.
My discussion paper, and all of my comments in this thread, do not presuppose any theory of design.
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 14:20
Yersinia,
Let me try coming at my question from another direction. Thought experiment.
Suppose (counterfactually) that when Smith et al. compared H. erythrogramma and H. tuberculata, the DNA comparison turned out to be no better than between any two randomly-selected sequences. (This is, I think, what you said earlier would be hard or impossible to fit with a hypothesis of common ancestry of the two species.)
But I say, "Ah, doesn't matter -- in gross morphology, behavior, and other respects, the two species still belong in the same genus. After all, that's where they classified before we looked at their DNA. We'll sort out the nucleic acid incongruities later."
Now it's up to you to persuade me that my hypothesis of common ancestry within Heliocidaris is wrong, because it violates the test (for common descent) that you proposed. How would you go about making your case?
I realize this post may cross (in time) with your latest reply, but hang in there and I'll get back to you.
[ 11. February 2003, 14:22: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
Josh
Member
Member # 405
|
posted 11. February 2003 14:21
I thought this reference might be of some use to Paul in his analysis of Generative Entrenchment and Ontogenic Depth.
"Systematic Functional Analysis of the Caenorhabditis elegans Genome Using RNAi." Nature 421, p231-237.
quote:
;"A principal challenge currently facing biologists is how to connect the complete DNA sequence of an organism to its development and behaviour. Large-scale targeted-deletions have been successful in defining gene functions in the single-celled yeast Saccharomyces cerevisiae, but comparable analyses have yet to be performed in an animal. Here we describe the use of RNA interference to inhibit the function of 86% of the 19,427 predicted genes of C. elegans. We identified mutant phenotypes for 1,722 genes, about two-thirds of which were not previously associated with a phenotype. We find that genes of similar functions are clustered in distinct, multi-megabase regions of individual chromosomes; genes in these regions tend to share transcriptional profiles. Our resulting data set and reusable RNAi library of 16,757 bacterial clones will facilitate systematic analyses of the connections among gene sequence, chromosomal location and gene function in C. elegans."
Interestingly, the authors group RNAi phenotypes into non-viable, growth defects, viable post-embryonic phenotypes and show their distribution on each chromosome arm, and additionally show the distribution of functional classes of genes within each category. They also look to see if phenotypes that would be worm-specific (i.e. viable post-embryonic phenotypes similar to losing an arm instead of disrupting embryonic patterning) are enriched in genes found only in worms VS distribution of genes related to other organisms in the subset of genes that give phenotypes that are required for basic development and viability (embryonic lethal.)
This may be an important direction in analyzing the hypothesis you are developing, with respect to the genetic differences between organisms that result in altered development, cellular organization, etc. As this type of screen is expanded to other organisms (they comment on the utility of doing the same analysis in a closely related nematode, C. briggsae) and the types of phenotypes are more carefully analyzed and grouped into much more specific classifications (like construction of the pharynx vs. heart development for early lethal phenotypes) one could see that ontogenic depth will be very well characterized. One could also imagine dissecting the "genes are everything" question by looking closely at proper gradient formation and organization of the oocyte, etc. This analysis will also address the Volvox issue raised earlier, as genes that are highly conserved between Volvox and C. elegans would be hypothetically more essential to C. elegans development than genes specific to C. elegans. For example, classes of genes that disrupt embryonic division patterning would be much more conserved among multicellular organisms than genes required for proper heart development (if basic embryological patterning and divisions are simply extensions of the mechanisms simpler multicellular species use for tissue organization.) I would guess that if one could perform RNAi in the closely related Heliocidaris species and perform this same analysis, some of the questions about common descent and mechanisms for developmental differences would be addressed.
[ 11. February 2003, 14:27: Message edited by: Josh ]
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 11. February 2003 14:36
quote:
Of course, but we'd also say why common descent would be falsified. That's the missing piece, which I'm hoping you'll provide (so that we can get back to the details of Heliocidaris).
What is it about the pattern you describe that would be impossible to reconcile with common descent?
Paul, how long is this going to continue? How much clearer can things get than "If process X was operating then we expect to see X', but we see Y' instead so the Y hypothesis is more supported than X by this particular dataset".
IP: Logged
|
|
yersinia
Member
Member # 324
|
posted 11. February 2003 14:52
Paul, in reply to your (latest) question, it would depend on what kind of DNA we had sequenced. For all I know, some non-coding junk DNA might diverge to randomness very quickly, and thus would not be a good test (just like radiocarbon dating is not a good test of the age of the earth past 50,000 years).
However, if the rRNA genes gave us this pattern when comparing Heliocidaris species, it would give a me a very big "whaaaa?". If many genes gave the same unexpected patterns then we would conclude that something very different than common descent had happened.
Let's not be naive falsificationists -- obviously the overriding concern is where the preponderance of the data points on any given question. You apply many tests and see if each one weakens or strengthens the hypothesis, and then decide what your confidence is in the hypothesis based on that.
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 11. February 2003 15:49
Paul writes about a designer being able to bypass the problems of GE:
"Paul has not described it, because absolutely nothing (including GE) "prevents a designer from doing whatever he pleases.
My discussion paper, and all of my comments in this thread, do not presuppose any theory of design."
Excellent clarification.
The term, "common descent", as used by Paul, is conflated with "naturalistic evolution without the addition of designer-provided information either via front-loading or at other times during the development of a species". We've already established that one can evaluate common descent with regard to a particular set of organisms without necessarily having to establish that *all life* might be related by common descent. That's good too.
Now we've established that GE observations are not incompatible with "common descent", in the more generic sense of the term which could include a theory of design. It took a while, but it happened.
Progress.
Now, what prevents a reasonable ID biologist from concluding "common descent" in the case of these sea urchins? Or perhaps to put it another way: If one presumes design, what data would one look to in order to determine whether two organisms might be related by common descent? I suspect the methods would be almost exactly like what Yersinia et al. have already proposed.
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 11. February 2003 17:12
Argon wrote:
quote:
Now, what prevents a reasonable ID biologist from concluding "common descent" in the case of these sea urchins?
Nothing. But the ID biologist would face all the same difficulties as his naturalistic colleagues. Absent a specific proposal, simply saying "design is possible" makes no difference to how the history of Heliocidaris is evaluated (which is why Yersinia's earlier comments about baraminologists, partial hybridization, and whatnot, are irrelevant).
Argon wrote:
quote:
Now we've established that GE observations are not incompatible with "common descent", in the more generic sense of the term which could include a theory of design.
But nothing is incompatible with such a theory. GE observations do challenge good old ordinary common descent, which is a scientific theory that makes testable predictions.
Any possible observation is compatible with descent plus design to get over the hard patches, which is the reason Darwin (rightly) regarded such ideas with contempt. Show me a theory of common descent plus design that makes specific predictions about whether GE should hold, or not, and then we'll have something to talk about.
[ 11. February 2003, 17:31: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
|
Topic Closed
Printer-friendly view of this topic
