peter writes:
PB: this is GUToB rule #3
Hi Peter. I'm afraid I don't have time to go hunting for whatever the GUToB is. It seems to me simply that you could address things in terms of current molecular knowledge -- especially as it applies to salty's claims -- and just leave it at that. But since it seems that you've been given a vacation, I'll just address a few points quickly.
PB: The genes you call unique have been derived from preexisting DNA elements, as predicted by GUToB.
Well
of course they've been derived from existing DNA elements. They are the result of gene duplication and the incorporation of non-coding elements into coding regions, among other things. The point is simply that they represent an increase in information via natural means. Salty has claimed that there hasn't been any new information. And the rest of the ID movement claims it couldn't have happened by natural means. They're both wrong.
In addition, it is highly unlikely that such genes arrived by random muations and selection (one can calculate a bit on it) and thus I advocate a non-random mechanism.
LOL! It
has been calculated, and they're perfectly capable of arising through random mutation and selection. Just read some of the papers.
Indeed, mechanism for non-random mutations (NRM) have recently been described. For an overview of NRM see Caporale's book "Darwin in the genome" and (almost) all my threads.
From reading reviews of Caporale, it doesn't appear that she advocates any non-random (with respect to fitness) mutational mechanisms. Instead she points out mutational mechanisms that are more likely to produce an adaptive effect, rather than those that "intentionally" strive for a particular result. It shouldn't come as a surprise from a Darwinian standpoint that molecular mechanisms that are more likely to produce good phenotypes will themselves spread. But that's something entirely different from a mechanism that intentionally hits a prespecified target in response to certain evironmental cues. Corporale appears to be every bit as Darwinist as myself.
The only such "adaptive mutation" that's ever been seen to occur is that from the work of Barry Hall. It was thought for a time that these might indeed be some sort of Lamarkian evolution, but it's since been discovered that they're fully Darwinian. I'm afraid there are no known mechanisms of the kind you seem to be alluding to.
You didn't get it. Functional DNA elements do not just drop out of the sky (that would falsify the GUToB). The 'novel' genes have been derived from preexisting DNA elements and are most likley re-assembled through NRM. Probably, there are gene-generating mechanisms in the genome.
Well
of course they don't drop out of the sky. Are you not aware of existing models for gene evolution? Most of them rely on duplication of existing elements, because that allows for redundancy, which allows for new functions to be added while old ones are retained. And then mechanisms like exon-suffling and transposition help create new genes by mixing and matching functional domains. The point is simply that they add "information" and complexity to the genome, contrary to salty's claims. Whatever it is that you're advocating, it doesn't seem any different than the standard mainstream evolutionary view, other than the fact that you're tossing "non-random" mutation in there for no good reason.
And if you want an example of a gene that does appear to "drop out of the sky", check out turf-13. It's a functional gene that came from entirely non-coding sequences.
PB: If the genes are really novel you would be right. But they are not. They are derived from preexisting DNA elements.
How are you defining "novel". Given that it's not really possible for a gene to come from something that
wasn't an existing DNA element, this is not a useful meaning of the term. What are you looking for, some DNA that forms
de novo from carbon, oxygen, and nitrogen? The whole point of evolution is that DNA comes from existing DNA, but that it can change or duplicate, resulting in creatures that are different than the DNA that they came from. You could keep moving the goal posts all the way back to the origin of life, but this is irrelevant to our current discussion (which incidentally, salty seems to have limited only to metazoans).
I'm using the term "novel" to mean a new and unique function, one that was not carried out by the DNA element from which the new gene came. They examples I provided and many others show that this can and does happen quite a bit. That's enough to falsify salty's claim to "no new information". Not to mention the core claim of the ID movement.
The recent origin of such genes make it highly doubtful that they arose through a random mechanism. For instance the gene in the LCR16a segment (Johnson et al, Nature 2001, 413;514-19).
That paper doesn't support your claims at all. The copies of this gene share 98% sequnce identity. If you want to find evidence against mutation and selection, find a gene of recent origin that has no homology to any known gene or non-coding squence. As it is, the authors drew up an nice little phylogenetic tree showing this gene's origin, and that of its paralogues in other apes.
And there is not the slightest hint of evidence that this gene required "non-random" mutation. You are probably talking about the fact that the substitution rate for the coding sequences was higher than that of the neutral sequences. Well guess what? That's what's expected from positive Darwinian selection. And that's exactly what the authors say. They were measuring positive selection, not the mutation rate. And you might want to check out their
recent paper for a model of the evolution of this gene family.
PB: The duplication-divergence mechanism sounds reasonable but cannot be the origin of the members of the Src-phosphatase family: point mutations give non-viable phenotypes (while knockouts are viable). In fact this family falsifies the evolutionary vision.
Cite please. There are an awful lot of Src proteins. Which protein(s) was knocked-out? On which protein(s) were mutations performed? Did they do saturation mutagenesis, or just mutations on the catalytic site? Based on your misunderstanding of the previous paper (and the one below), I strongly doubt that the literature on Src supports your claims in any fashion.
PB: My still unaddressed question I've asked almost a year ago: When a functional redundant gene is knocked out is there a loss of information?
Depends on how you define information. In most cases, yes.
It should be noted that redundant genes do not have an association with gene duplication and do not mutate faster. (in: A genetic uncertainty problem by D. Tautz, TiG 2000; 16:475-7)
Did you read further than the first two sentences of the abstract?
quote:
In Drosophila, for example, many of the early segmentation genes have partially overlapping functions and, although they are members of different gene families, can replace each other for certain regulatory aspects[4]. This suggests that there could be a positive advantage for selecting this type of redundancy during evolution, such as higher fidelity within the regulatory or developmental network [4]. Explicit modelling of this assumption does indeed show that unrelated genes are likely to be recruited to complex pathways during the course of evolution [5]. But the alternative `robustness against mutations' hypothesis can also be modelled and leads to the evolution and maintenance of redundancy under realistic population parameters [6].
Indeed, the
point of this paper is to explain in evolutionary terms why these redundancies exist. By making it look like an unsolved mystery, you are giving people the impression that this is somehow an uncrossable barrier for evolution. You are using an old creationist tactic, and discrediting yourself in the process.
Furthermore, if you know anything about knock-outs, you'll realize why they're not always a good way to detect function. Laboratory organisms are kept under strictly controled conditions, such that any subtle effects which might be important in the wild will not matter in the lab. Knock-outs are best for detecting essential genes or those that affect gross morphology, but are limited in detecting other functions. The paper you cite makes that exact point too:
quote:
However, invoking redundancy is not the only possible explanation for gene knockouts that have no phenotype. One could simply propose that it is just a question of finding the right experimental conditions to detect an effect. As Louis Wolpert once remarked about no-phenotype knockout mice, `But did you take it to the opera?' This is, of course, a valid question and it suggests that there might be almost no limit to possible experimental efforts. By contrast, one could ask whether it is possible that a gene could have such a small function (or fitness benefit) that it could only be revealed by such a tremendously large experimental effort?
In fact, this is exactly what population genetical theory predicts. Kimura's theory of neutral evolution shows that, in large populations, vanishingly small selection coefficients are sufficient to fix a gene that conveys a selective advantage.
[...]
Are there any data that suggest that weak selection effects do indeed shape genes and genomes? The no-phenotype knockouts could be taken as such evidence, but, as discussed above, they could also be explained by invoking redundancy (although the redundancy concept also operates with weak selection). More direct evidence has been presented by Thatcher et al.[11], using the theoretical framework of neutral evolutionary theory outlined above to test whether genes that contribute marginally to fitness exist in yeast. Using competition experiments that can trace even very weak differential fitness benefits in yeast, they found that 70% of the nonessential genes tested resulted in a reduced fitness under appropriate competitive experimental conditions, a result that is in line with the comparable study by Smith et al.[2].
Analysis of the patterns of codon usage patterns also shows not only that weak selection operates, but also that the genetic formula for the general population is valid.
Good grief man, why are you asking
me to explain this for you when the paper you cite has the answers?
PB: Apparently biological information can not be defined that way, since non-phenotype knockouts are rather the rule than exceptions.
But did you take them to the opera?
The examples I cited have an obvious function. For some of them, the function is still unknown. But in any case where you have a biological function being determined by a DNA sequence, you've got information as far as I'm concerned.
PB: What is the selective constraint to keep redundancies stable in the genome?
Try reading the paper you cited. It answers that question quite nicely.
[This message has been edited by Grape Ape, 03-21-2003]
™ Version 4.2