Macroevolution: Its all around us...

Hi randman,I am hopping in late here, so if you get a chance would you please specifically state your "assessment of convergent tendencies within DNA" so I can see where I stand on that assessment?Thanks for citing the Vowles and Amos work - it is definitely interesting, and points out that the concept of "random mutation" is a simplified one. The molecular nature of some DNA sequences, particularly repetitive ones, make them more mutation prone than non-repetitive sequence. It appears that this bias extends to regions flanking certain types of repetitive sequence as well (or perhaps that repetitive sequence stabilizes certain flanking sequences thereby providing selection).However, it seems to me (please correct me if I'm wrong) that you are asserting that: because convergent evolution is apparent in certain non-coding sequences, sequence similarities cannot be used to establish common ancestry.That's an overreaching jump in logic - it may be that we simply cannot used those specific types of non-coding sequences in isolation as a way to determine the nature of common ancestry. However, we still have the entire non-repetitive genome to use to establish common ancestry, since there is no evidence for convergent evolution of such sequences. The study you cite doesn't have an impact on using pseudogene sequence to establish ancestry, for example.

First - I don't think anyone is claiming that similar traits can only be explained by common ancestry.Second - I feel there is a strong understanding of DNA sequence evolution. There is no evidence for convergent evolution of coding or pseudo-coding sequence. Take the popular GLO gene as an example (it is involved in vitamin C production). The gene is mutated so that it is non-functional in humans, chimps, fruit bats, and guinea pigs; thus these species must get vitamin C from their diet.Humans and chimps have the exact same mutation in the GLO gene, while the guinea pigs and fruit bats have different, distinct mutations. The current theory is that there were three mutational events leading to the three 'broken' genes in the four species - and that the GLO gene was mutated in a common ancestor of humans and chimps, with both species inheriting the mutation.Does convergent evolution make sense in this case? How would you test it?It doesn't make sense to me that the arisal of identically not-quite-functional-genes would arise by convergent evolution in two species, especially since there are countless examples of similar cases throughout the genomes of chimps and humans.To put it slightly differently, why would all of the sequences in the human and chimp genomes end up more closely convergently evolved than, say, the mouse genome? If sequences were converging to a particular sequence in multiple species, and you compared numerous sequences across those species, there should be no correlation to morphology - instead there should just be genetic "noise". If you haven't figured it out yet, science is tentative, so it is always the case that "the best we can say is maybe". Science proceeds via falsification - common ancestry has yet to be falsified by any data, and there is an unimaginably enormous amount of data that tests common ancestry. I don't disagree, but you have to be careful here in switching between shared traits and shared sequence. The truth of the matter is that shared traits and shared sequence produce matching "family trees", and importantly, the result is the same even when sequence that does not contribute to phenotypic traits is used. But is it convergent evolution at the DNA level at all reasonable to explain this commonality? Is it reasonable that all shared genes across all species convergently evolved? (For example, that the Cytochrome C gene arose separately in every single species on the planet, from bacteria to badger?)That seems like an amazingly far-fetched explanation to me... and combined with the other characteristics of sequences I describe above, I think we can logically cross "convergent evolution" and "common design embedded in the universe manifesting in DNA convergency" off the list as being solely responsible for the commonality.And an interesting side note: What I think is really fantastic is that your "common design embedded in the universe manifesting in DNA convergency" is essentially a rewriting of Haeckel's "biogenetic law" at the genetic level instead of the phenotypic level. That is the same law that lead him to start fradulantly representing embryos in the diagrams whose use you so abhor. It is also a "law" that has been continually falsified since the 1870s.

Randman-First off, stop being so reactionary, and please take the time to read my posts. You state: I NEVER denied convergent evolution of DNA. I support it, at least for the types of DNA sequence that there is evidence for convergent evolution. I simply feel that your extension of the specific evidence you have to theories of common ancestry and evolution in general is overreaching. It seems that rather than deal with the actual criticism I made you'd accuse me of denying evidence, evidence that I supported and even thanked you for bringing to the discussion. Why? What is the specific prediction that "special creation theory" makes that suggests that the results of convergent evolution would produce closer sequence similarities between chimp and human than between human and mouse?You can't just say "the data fits special creation perfectly". You have to explain how and why the data fits special creation, and how special creation predicted that data trend. Is it really "an area of intense study"? How many labs are working on the problem? How many papers have been published on the subject?You've provided a great piece of evidence NOT that non-coding DNA exhibits convergent tendencies, but that non-coding DNA flanking specific types of microsatellite repeats exhibit convergent tendencies. Do you understand the difference? I never suggested you argued a sole explanation. Again, perhaps you should take more time reading posts before responding to them. In fact, I acknowledged that you listed several commonalities. Are you interested in having a decent discussion, or just pointing accusatory fingers? No, my line of reasoning rests on the fact that different types of DNA sequences evolved differently. You can't simply take evidence from one class of DNA sequence and immediately extend it DNA in general - that is where I believe the root of your "overreaching" lies.And there IS much else to say, since you by saying I am incorrect you are asserting that there is, indeed, evidence for convergent evolution of coding or pseudo-coding genes.Please provide that evidence (hopefully you realize that the reference you've already given does not give such evidence).

Hi WK-Doesn't "highly unlikely" seem a bit strong, especially since the paper is examining such short repeats? As you mention, (AC)2 repeats produce significant influence (and the influence seems to increase greatly at (AC)5) - is there any reason they wouldn't be in the coding sequence of the gene? There are plenty of genes with short microsatellites within their coding sequences, even more with microsatellites in introns, no?Here is an interesting paper on the subject (more or less) from a few months ago (though it did not focus on AC micosatellites, and actually analyzed sequence at the amino-acid level): What are your thoughts?

Thanks for the apology - I do appreciate it (one of pet peeves is when people claim I made arguments that I did not). I guess I'm not sure precisely what you mean here. At the non-sequence level DNA is essentially the same thing. BUT, we are not discussing non-sequence level behavior of DNA here, we are discussing the influence of specific types of sequence, and thus the sequence level behavior of DNA. Thus we cannot simply say "all DNA is the same, therefore it must behave the same way", because it is not all the same.Do you see the difference?Just because AC repeats appear to induce or stabilize convergent evolution in nearby DNA in no way implies that the sequence of the Cytochrome C gene, for example, has undergone convergent evolution.Another way of putting it; repetitive stretches of DNA sequence have a higher rate of mutation than non-repetitive DNA sequence. Thus even though at the non-sequence level it is "all the same DNA", the sequence does indeed influence the biochemistry of the DNA.Hopefully that all makes sense...If it makes sense maybe you'll see why I think your statement- -is a simplification that leads to that leads to overreaching the evidence. That would be the case if any pattern was indicative of design, which I do not feel is the case. Just because certain DNA sequences form a pattern because they are mutable or stabilized at a molecular/biochemical level does not indicate design to me, especially since those sequence have no apparent function outside of themselves.In other words, I see chemistry as the result of the pattern where you see design - when minerals stabilize themselves into patterned crystals, I see chemistry at work, not design. What about you?

The first part of my post was the part most pertinent to this discussion.You can't dismiss my comments by simply claiming, "I think there is more evidence out there that agrees with me."Provide the evidence. Don't just tell me it might exist.
_______________________________________As a note, from the Vowles and Amos paper:

​

To our knowledge, no one has yet conducted a systematic study of mutational biases operating around microsatellites.

​

It appears they were the first to conduct an in-depth study of the problem. Neither author has published on the subject since then. And according to the ISI Web of Science the article has not yet been cited by another article. Since it's only been a year, this is not a great surprise, but it does suggest that there isn't a large community studying the problem...

I'm now wondering about the strength of this form of "convergent evolution" - in my mind I think of it as a rather "weak" force that would not be able to overcome the "strong" selective force of maintaining a phenotypically beneficial gene or sequence.Also, it seems to me that the convergent pattern they suggest would be unlikely to converge to functional gene sequence; that is, the pattern they uncover is unlike the vast majority of gene sequences. Do you think that is a fair conclusion?

In the specific case of dinucleotide repeats analyzed in the article, it is quite likely that they arose independently, and thus do not need "seeds" (though once formed they are heritable to "daughter species").This is the case because dinucleotide repeats are hypermutable, and are especially good at expanding and contracting. Keep in mind that in the paper WK referenced, the sequence:ACACwas all that was required to significantly influence mutational bias in flanking sequences - a "seed" sequence is not needed to explain four bases. If that sequence expands to:ACACACACACthe influence is much greater, and moreso at (AC)9, etc...Thus using sequence in or around such repeats could create confounding results. What randman doesn't seem to have grasped yet is that the rest of the genome remains to be used to establish common ancestry, since it doesn't appear that the convergent mutational bias is a particular issue for other types of sequence. Also, from my reading of the paper the flanking sequence is not converging to specific non-repetitive sequence, but rather is developing a simple pattern, and may indeed be converging to simple repetitive sequence itself...

I think I've seen you yourself post descriptions of how repeats allow "slippage" errors during replication to expand or contract repeats.This is new because it is quite different from that mode of mutation - the paper describes such repeats biasing mutation in non-repetitive flanking sequence at some distance, separate of slippage, so that a sequence like (repeat in red):GTGACGGGTACGTACACACACACACACGCGCTATATAGCwould converge via biased mutation or stabilization to something like:GAGACAGATACATACACACACACACACGAGATATATAGA(Someone please correct me if I misread the paper.)See how a pattern is forming, not due to simple slippage-based expansion/contraction? That is what is novel about the results...This message has been edited by the sasquatch aquatic, 06-23-2005 06:06 PM

No shit. You don't seem to have very good reading comprehension. What do you think the phrases "I'm now wondering..." and "in my mind I think of it as..." mean?It's called speculation, not assertion. It doesn't assume that at all. You really need to take time and think about posts rather than quickly try to twist them into something you can take argument with.In no way did I state that they were necessarily opposite. Obviously two forces in the same direction would be additive - but that doesn't allow us to think about relative strength of those two forces, which is the nature of my conjecture.Which is: Mutational bias or stabilization is a weaker selective force than natural selection acting on phenotypic traits, since phenotypic traits effect fitness and survival more than mutation of a single base in non-coding sequence near untranslated dinucleotide repeats. You judge it to not be a fair conclusion when you're not clear what I was talking about? That's fair.From my reading of the paper, the flanking sequence was mutationally biased towards repetitive sequence. Gene coding sequences don't contain that degree of repetitive sequence. Therefore, it appears that the convergent evolution of the flanking sequences was actually making said sequences less gene-like. I was asking WK if he though my reading of this was correct. If they were equally positive they would be additive or cancel each other out. One would not be dominant.

There is no evidence that "DNA in general is convergent in non-coding sequences". That requires a huge leap of faith.(Of course preceding it with "the likelihood exists" renders the statement rather meaningless. The likelihood exists that I am an invisible ninja that can stop the rotation of the planet with my mind.)

Two guys arm wrestle. One guy pushes first, but is much, much weaker, and the other guy beats him easily.Your "first come, first served" idea doesn't make any sense. Sorry, incoherent, and from what I can tell, based on serious assumptions. Right - you're actually restating my conjecture that you said was incorrect a few replies back. Also, I've mentioned many times that repetitive sequences may "stabilize" certain sequences in flanking regions. That is selection at the molecular genetic, rather than phenotypic level.And has nothing to do with convergent evolution of genes at the level of mutational bias or stabilization, which is what you have been trying to argue.

randman, you simply do not understand how science proceeds: Science cannot falsify a negative; in fact, science cannot prove anything.Science can only test falsifiable theories, and either support or falsify them. I'm not even sure what you mean by "DNA convergency" anymore. Do you mean DNA sequence undergoing convergent evolution within a genome (as described in the V&A paper); or convergent evolution of genes between species? We currently have no evidence that DNA is inherently undergoing convergent evolution.We only have evidence that one specific type of DNA sequence influences the evolution of flanking sequence.There is no reason to believe that DNA in general (particularly "functional" sequence) is undergoing convergent evolution.I'll wait for the evidence before I make the overreaching leap that a trend found in one very specific type of sequence applies to all sequence. No, that is on par for the comments you have made throughout this thread. You repeatedly state that "DNA is convergent" as fact when it is simply a huge, unreasonable assumption.

If they are not adversarial, then they are working in the same direction. Therefore, it doesn't matter who gets there first, the result will be the same. You have stopped making any sense. I accept your analogy, but not your conclusion.Because in order for your analogy to be correct the race would have to be a few million years long. Plenty of time for the fastest man to be born, train, and beat your out-of-shape ass... For that to be the case you'd have to assume that non-biased mutations don't occur. Is that what you are arguing? I already explained that I was discussing relative "strengths", and which would overpower the other if they were in opposition. I already explained that they aren't necessarily adversarial.Again, stop being so reactionary, and actually read the posts you are responding to.

No. You cannot extend the results of this study to all DNA.As an example, there are many studies showing that certain DNA sequences code for protein sequences. By your logic, we could assume that all DNA codes for protein, which we know is not the case.Alos, it has been know for quite some time that DNA mutation is non-random, long before the study under discussion.