|
Author
|
Topic: The GUToB
|
Pim van Meurs
Member
Member # 541
|
posted 23. April 2003 17:40
More on randomness as used in (neo)-Darwinism, thanks to Mark24
quote:
Mutation as a Random Process
Mutations occur at random. It is extremely important to understand what this statement does & does not mean. It does not mean that all conceivable mutations are equally likely to occur, because, as we have noted, the developmental foundations for some imaginable transformations do not exist. It does not mean that all loci, or regions within a locus, are equally mutable, for geneticists have described differences in mutation rates, at both the phenotypic & molecular levels, among & within loci (Woodruff et al. 1983; Wolf et al. 1989). It does not mean that environmental factors cannot influence mutation rates: ultraviolet & other radiation, as well as various chemical mutagens & poor nutrition, do indeed increase rates of mutation.
Mutation is random in two senses. First, although we may be able to predict the probability that a certain mutation will occur, we cannot predict which of a large number of gene copies will undergo the mutation. The spontaneous process of mutation is stochastic rather than deterministic. Second, and more importantly, mutation is random in the sense that the chance that a particular mutation will occur is not influenced by whether or not the organism is in an environment in which that mutation would be advantageous
(Evolutionary Biology 2nd Edition, Douglas Futuyma p281-2)
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 00:39
Hi Pim,
quote: "Mutation is random in two senses. First, although we may be able to predict the probability that a certain mutation will occur, we cannot predict which of a large number of gene copies will undergo the mutation. The spontaneous process of mutation is stochastic rather than deterministic."
I know it is not new that mutations can be introduced independently on the same spot. It has been demonstrated over and over. What I am trying to convey is that we have underestimated the effects of NRM in phylogenetic analysis: they line up and give an illusion of common descent. This was my point in the ZFY example. It was very clear from the sequences that the ZFY region does NOT demonstrate common descent but rather a common mechanism that gives the illusion of common descent.
Now, it is up to the evolutionists whether or not they are going to exclude the mechanisms of NRM1 and NRM2 when they do phylogenetic analyses. The ZFY region demonstrates that it is very tricky to rely upon only one sequence per species. For instance the first chimp sequence demonstrates point mutations that would not be expected from common descent since they are only shared with Lemur and Tamarin. I think we do not have a molecular evolutionary explanation for this observation. Furthermore, I have noticed that such strange alignments of mutations are becoming visable and are common as soon as subspecies are analysed and presented.
For instance the Drosophila subspecies. From the 1g5 gene it is clear that the mutations in this NEUTRAL region are introduced by a mechanism, since selection on these neutral positions can be excluded. Similar results have been described for mutations in T4. Linn Ripley was even able to PREDICT where the mutations were going to be introduced. If one can predict WHERE mutations are going to be introduced, then such mutations are not introduced randomly. Something similar is going on in human mtDNA. What we need is more comparisons of subspecies to elucidate the underlying mechanism. It also explains why mtDNA is not a good tool for a molecular clock.
quote:
"Second, and more importantly, mutation is random in the sense that the chance that a particular mutation will occur is not influenced by whether or not the organism is in an environment in which that mutation would be advantageous".
I have already mentioned that this is very questionable for the toxin genes in Cone snails (reference to in previous mail). The mutaions in the toxin genes are for sure advantageous in the environment where the ofspring of the snail is produced.
Taken together, if we know the underlying mechanism mutations can be predicted and surely some mutations are deliberately introduced in genes through a preexisting mechanism in a gene that determined succes in a particular environment. I see two little problems for NDT here.
best wishes,
PB
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. April 2003 01:07
I'm really unclear on what the 1g5 data is supposed to show.
First, what is NDT? Neo-Darwinian Theory? Neutral Drift Theory?
That there are hot spots, cold spots, grouped mutations by excision repair, and so on, is hardly controversial. If you are only arguing that mutation is not a uniformly-at-random selection of nucleotides across the genome, you are preaching to the choir! Of course, people doing sequence alignments don't like to take all of these factors fully into account because it makes alignment much harder--but that doesn't mean that there's a crisis for Darwinism, just that reality is messy and, as usual, we try to ignore some of the mess when we can, and use independent evidence to confirm that the ignored mess wasn't too important.
In terms of common descent vs. repeated mutation, you haven't provided any evidence that the genetic similarities of geographically separated populations aren't due to migration. Flies move. Humans have spread across the planet several times. Furthermore, you have picked out Australia, Russia, and Canada, but at least in the table you gave there, the same sequence is also shared by flies in Iraq, Cyprus, and the U.S.--and appears to be the original sequence, given the figure you linked to. So the pattern doesn't seem all <i>that</i> peculiar. And also, there's no data on how uniform the sequence is within a population. There are apparently at least three different U.S. and Australia sequences (each).
And while it is possible that the percentage differences between introns reflect different mutation rates, it is also possible that D. simulans subtypes have existed independently for longer. How can you differentiate between those two explanations for the data?
If the authors claim that the whole region is undergoing neutral evolution and you say it is not, then you have to consider both that (1) there is selection and (2) your active mechanism is operating. How do you know that there isn't selection? Your observations are intended to undermine the data that they presumably used to decide whether there was selection or not.
With respect to Src and Hck, I'm not sure that the mechanisms of chromosomal rearrangement are perfectly understood. But even so, being "stringently regulated" and "involving several recombinational proteins" doesn't equal (completely) "non-random". For example, UV or chemically induced double-strand breaks typically induce emergency repair mechanisms ("stick the free DNA ends together!"); this repair requires protein machinery, yet that machinery doesn't specify where the break occurred.
And the rest of this is very long. I'm sorry, but I wanted to show my work so that it was apparent that this wasn't just opinion. Hopefully I've included enough detail to make what I've done reproducible.
With respect to the ZFY sequence, I am not sure how you escape the conclusion of common descent. Let's list the variable bases (noting the failure of the HTML to line up the way it should):
code:
Hsy Pt1 Pt2 Ggo Ppy Hag Mfu Cae Lca Str
# 1: G G G G G G G G G T
# 21: C C C C C C C C T C
# 25: C T T T T C T T C T
# 31: T C T T T T T T T T
# 52: T C T T T T T T C C
# 58: T T T T T T T T C C
# 61: G A A A G G G G T G
# 76: T T T T T T T T C T
# 91: C T T C T C C C T T
#103: G A G G G G A A A A
#106: C C C C T C C C C C
#107: A A A A G A A A A A
#115: G G G G G G G G A G
#119: C C C C C C C C T C
#139: C C T T C C C C C C
#154: G A G G G G A A A A
#163: G G G G G G G G A G
#172: C C C C C C C C T C
#175: C C C C C C C C T C
#178: G G G G G G G G A G
#181: C G A A A C G G T G
#184: C C C C C C C C C A
That's halfway through, and I am now bored with typing these in. Let's construct a table of differences
code:
Hsy Pt1 Pt2 Ggo Ppy Hag Mfu Cae Lca Str
Hsy 0 8 5 4 5 0 4 4 15 7
Pt1 0 5 7 8 8 4 4 14 9
Pt2 0 1 4 5 6 6 17 9
Ggo 0 5 4 5 5 17 10
Ppy 0 5 6 6 17 9
Hag 0 4 7 15 7
Mfu 0 0 15 5
Cae 0 15 5
Lca 0 13
Str 0
From this, we see that Hsy and Hag are identical (let's call them Hxx), as are Mfu and Cae (let's call them Mxx). We can also observe that Lca is way different from the others, and that Str is fairly different. Everything else is pretty similar, so that's about all we can get off of the first half of the table without trying to reconstruct a history.
So let's try to reconstruct a history. First we look at the sequences and throw out singletons since they won't tell us anything about who is related to whom:
code:
Hxx gcctttgtcgcagccggccgcc
Pt1 gctcctattacagccagccggc
Pt2 gcttttattgcagctggccgac
Ggo gcttttatcgcagctggccgac
Ppy gcttttgttgtggccggccgac
Mxx gcttttgtcacagccagccggc
Str tcttccgttacagccagccgga
Lca gtctcctctacaatcaattatc
xx xx x xxxx xxxx x ?singleton
Hxx c tg cg cg c
Pt1 t ta ta ca g
Pt2 t ta tg tg a
Ggo t ta cg tg a
Ppy t tg tg cg a
Mxx t tg ca ca g
Str t cg ta ca g
Lca c ct ta ca t
Hxx Pt1 Pt2 Ggo Ppy Mxx Str Lca
Hxx 0 6 4 4 3 4 6 6
Pt1 0 4 5 4 2 2 4
Pt2 0 1 2 6 7 7
Ggo 0 3 5 7 8
Ppy 0 4 4 6
Mxx 0 2 5
Str 0 3
Lca 0
Here we can identify two more blocks: Pt2/Ggo/Ppy are all quite similar, as are Pt1/Mxx/Str. Now we're ready to start constructing a hypothetical history.
code:
|---------- Hsy/Hag gcctttgtcgcagccggccgcc
| ^ ^^ ^
| |----- Mfu/Cae gcttTtgtCacagccagccggc
| |----- Pt1 gctCctAttacagccagccggc
| |----- Str TcttcCgttacagccagccggA
|----| AncestorA gcttctgttacagccagccggc
| ^ ^ ^ ^ ^
| |----- Pt2 gcttttattgcagctggccgac
| |----- Ggo gcttttatCgcagctggccgac
| |----- Ppy gcttttGttgTGgctggccgac
|----| AncestorB gcttttattgcagctggccgac
| ^ ^ ^
|----| AncestorC gcttttgttgcagccggccg?c
|
|--------------- Lca gtctcctctacaatcaattatc
|
----| AncestorD g??t????t?ca??c??????c
caps: mutation from first-tier ancestor
caret under a letter: mutation from second-tier ancestor
This is pretty good, but we can do a little better. If AncestorC has a T at position 5, then maybe Mfu/Cae is the ancestral one, rather than mutating and mutating back. (This is completely consistent with the tree.) Likewise, we can guess that it was G at position 7 originally, and that Pt2 and Ggo share the G->A mutation by lineage. So our refined tree looks like
code:
|----- Hsy ......................
|----- Hag ......................
|---------| Hxx gcctttgtcgcagccggccgcc
| ^ ^^ ^
|
| |----- Mfu ......................
| |----- Cae ......................
| |----| Mxx gcttttgtCacagccagccggc
| |
| | |----- Pt1 ...c..a...............
| | |----- Str t....c...............a
| |----| Sxx gcttCtgttacagccagccggc
| |
|----| AncA gcttttgttacagccagccggc
| ^ ^ ^ ^
|
| |----- Pt2 ......................
| |----- Ggo ........c.............
| |----| Gxx gcttttAttgcagctggccgac
| |
| |---------- Ppy gcttttgttgTGgctggccgac
| |
|----| AncB gcttttgttgcagctggccgac
| ^ ^
|
|----| AncC gcttttgttgcagccggccg?c
|
|
|-------------------- Lca gtctcctctacaatcaattatc
|
|
---| AncD g??t????t?ca??c??????c
Innermost difference: letter vs. dot
Mid-level difference: capitalized
Outermost difference: caret under letter
Whew! What a long-winded exercise in imagination, right? How could we possibly tell if this is correct?
Luckily, we have the other half of the sequence! If the hypothesis of common descent is correct, the other half of the sequence should produce the same tree. Alternatively, if we put the other half of the sequence in the same tree, it should fit well. If there is some functional similarity in mutations, there is no particular reason to expect the commonalities to hold, however. So while repeated mutation is consistent with tree sameness, the prediction from common descent is much stronger: the trees should be the same. (For a tree with 10 terminal nodes, there are something like a thousand different ways to draw it.)
If we restrict ourselves to the sequence that is different, in the second half of the sequence we have
code:
Hsy tgtgtcaaaacttcggtttaaa
Pt1 tgtgtttgagattgggcttgtg
Pt2 tgtatcaggactccagtttaag
Ggo tgtatcaggactccagtttaag
Ppy tgaaccaggactccagtttaag
Hag cgtatcaaaacttcggtttaaa
Mfu tgtgtctaagattgggcttatg
Cae tgtgtctaagattgggcttatg
Lca tatatttaagcctgggccagtg
Str tgtgtttaagattggacttgtg
And if we put this into our tree and work backwards--using the tree to decide what the ancestral mutations must have been--we get
code:
|----- Hsy ......................
|----- Hag c..a..................
|---------| Hxx tgtgtcaaaacttgggtttaaa
| ^
|
| |----- Mfu ......................
| |----- Cae ......................
| |----| Mxx tgtgtctaagattgggcttatg
| |
| | |----- Pt1 .......g..............
| | |----- Str ...............a......
| |----| Sxx tgtgtttaagattgggcttGtg
| |
|----| AncA tgtgtctaagattgggcttatg
| ^ ^^ ^ ^
|
| |----- Pt2 ......................
| |----- Ggo ......................
| |----| Gxx tgtatcaggactccagtttaag
| |
| |---------- Ppy tgAaCcaggactccagtttaag
| |
|----| AncB tgtatcaggactccagtttaag
| ^ ^^ ^^^
|
|----| AncC tgtgtcaaaacttgggtttaag
|
|
|-------------------- Lca tatatttaagcctgggccagtg
|
|
---| AncD t?t?t??aa?c?tggg?????g
Innermost difference: letter vs. dot
Mid-level difference: capitalized
Outermost difference: caret under letter
The fit is remarkably good. I don't see a better tree than the one I've drawn.
If you complain about the three-way split, there's a way to disambiguate that as well. I constructed Ancestor C by letting Hxx, AncA, and AncB cast one vote each--whichever one had two determined Ancestor C. But this is cheating, since *some* of the time, you'll get two votes because the common ancestor of a pair of them had the mutation, and the singleton stayed as the original. We can tell which was the case by comparing to Lca; here you find that Lca agrees with singleton AncA most of the time (6 times it agrees with A, as compared with 2 times for H and 1 time for B). So this means that AncA is probably the odd-one-out of the Hxx, AncA, AncB trio.
Putting this information back in gives us our final tree, which we can't verify because we're out of sequence:
code:
Only changes are noted!
|----- Mfu ............................................
|----- Cae ............................................
|----| Mxx ........C...................................
|
| |----- Pt1 ...c..a......................g..............
| |----- Str t....c...............a...............a......
|----| Sxx ....C......................T.............G..
|
|---------| AncA ....................g....g......a...........
| * * *
|
| |----- Hsy ............................................
| |----- Hag ......................c..a..................
| |---------| Hxx ..c.....c...........c....g.................a
| | ^ ^ ^ ^ ^
| |
| | |----- Pt2 ............................................
| | |----- Ggo ........c...................................
| | |----| Gxx ......A.....................................
| | |
| | |---------- Ppy ..........TG............A.C.................
| | |
| |----| AncB ..............t.....a........gg...cca.......
| | ^ ^ ^^ ^^^
| |
|----| AncC' .........g.....g............a..a......t...a.
| * * * * * *
|
|----| AncC gcttttgttacagccagccg?ctgtatctaagcttgggcttatg
|
|
|------------------------- Lca gtctcctctacaatcaattatctatatttaagcctgggccagtg
|
|
---| AncD g??t????taca??ca?????ct?tat?taagc?tgggc???tg
Number of mutations from AncC: 10111120121100110000311013111211101111100111
Multiple mutations same? ------X------------------X---X--------------
(X = same mutation happened twice)
Number of mutations from AncD: 11212231121111111111311113121211111111111211
Multiple muations same? --X-XXX------------------X-X-X-----------X--
Mutation type C<->T . .. .
Mutation type A<->G . . . .
Total number of transitions (C<->T, A<->G): 57
Total number of transversions (anything else): 9
Difference at most recent divergence: letter vs. dot
Difference at next most: capitalized
Difference at next most: caret under letter
Difference at least recent divergence: asterisk under letter
We finally have some real numbers we can chew on. As you can see, there do seem to be hotspots where we get two or three mutations. However, nothing too whacky is going on here, since every single one of these is multiple transitions, as you'd expect given the high frequency of transitions to transversions. More sophisticated scoring models take this into account in their nucleotide substitution frequency tables, so this won't dramatically alter conclusions about relatedness. (The issue of hotspots may mess up the accuracy of using base pair changes as molecular clocks, but it tends not to have a huge effect on the topology of phylogenetic trees since hotspots are random, just with a different frequency.)
The hotspots aren't *that* bad, though. With 44 sites mutated out of 400, you'd expect about 5 overlaps and about 1 triple overlap. We actually see 10 double overlaps and 3 triple overlaps.
Now, we just have one last thing to check: does this make sense at all given what we know about the histories of the organisms?
The closest clusters
- Mfu and Cae--two types of similar monkey; looks good.
- Hsy and Hag--two gibbons; good.
- Pt2 and Ggo--chimp and gorilla; good.
- Pt1 and Str--chimp and tamarin (?!?!?!) Weird!
- Gxx and Ppy--chimp/gorilla and orang; good.
- Mxx and Sxx--monkeys and tamarin/chimp; hard to evaluate with chimp (probably good with tamarin).
- AncB and Hxx--chimp/gorilla/orang and gibbons. Nicely clustering apes.
- AncC vs. Lca--lemurs nicely separated from monkeys/apes
So the molecular phylogeny, independent of any knowledge of the relationship of these animals from other methods, has given almost a perfect report.
The only thing that is bizarre is the supposedly second chimp sample of Ptr1.
If true, this is quite peculiar. However, everything else fits so well that I wouldn't be surprised if this were an error in data collection. (Alternatively, if this gene is duplicated several times, and we're looking at different ones throughout the tree, the whole analysis could be suspect; I don't really know, since I don't know who picked the data set and how careful they were. I *hope* they were honestly picking a sample, not cherrypicking.)
Personally, I doubt Pty1 is a chimp at all; somewhere along the line, it probably got mislabeled (or was deliberately mislabeled by the list-compiler in order to check whether anyone was actually looking at the data properly).
Anyway, bottom line is: modulo one questionable data point, common descent seems to be well-verified and the standard theory seems in excellent agreement with this data set.
If you disagree, please show how the data "lines up to give an illusion of common descent". That is, point to a clear indication that the tree above is an illusion.
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 24. April 2003 01:11
Peter: This was my point in the ZFY example. It was very clear from the sequences that the ZFY region does NOT demonstrate common descent but rather a common mechanism that gives the illusion of common descent.
An interesting assertion that seems to lack only in evidence. In fact as I have shown the ZFY and other gene sequences very well match the fossilized evidence of primates for instance. All you did was to look at some regions and make some unsupportable assertions about their nature. I am not sure what this was supposed to prove but it surely did not demonstrate the presence of a "common mechanism that gives the illusion of common descent".
It is very well known that gene trees and species trees need not be the same due to the various genetic processes which can complicate the tree. That however does not mean that it is impossible to derive consensus trees based on more data.
Your claims such as "For instance the first chimp sequence demonstrates point mutations that would not be expected from common descent since they are only shared with Lemur and Tamarin. " seems to suffer from several problems namely that they were not be expected from common descent. Other than an assertion I have seen nothing that proves this point or even hints at such but once you provide us with your arguments perhaps things will look differently, I dont' know.
Despite the various references that show that Neo-Darwinism does not depend on randomness of position or timing of mutations but merely on randomness of the mutation wrt the environment. Your comments about the cone snail not only seem to contradict the comments by Caporale on this topic but also do not go beyond an unsupported claim by stating "The mutaions in the toxin genes are for sure advantageous in the environment where the ofspring of the snail is produced", a comment which surely ignores any attempt to provide such evidence other than through pure assertion. Are you sure for instance that selection does not play a role at all in this?
So despite all your claims about problems for NDT you have failed to provide any evidence that would support such claims.
I suggest that you present us with your arguments in a coherent manner so far the snippets of claims which lack in much supporting evidence make it hard to have a meaningful discussion about your ideas other than to correct many of your claims.
Some useful references to the cone snail may be helpful to understand that selection seems to have played a significant role here.
For instance
quote:
The answer may lie in the rather difficult way Conus snails manage to survive. Not the sort of creatures that can chase down their prey, the snails probably have to continually upgrade their arsenal of conotoxins to deal with changes in their prey. Some prey may develop an immunity to some toxins, or maybe new prey replaces the creatures that the snails had been dining on.
“There may be an arms race going on between the predator and the prey,” Duda says.
As a result, Conus snails have developed a versatile warehouse of toxins that can work in many different ways on different types of prey.
Source
The following paper which I have yet to read sounds at least interesting: Duda, T.F. and S.R. Palumbi. 2000. Evolutionary diversification of multigene families: Allelic selection of toxins in predatory cone snails. Molecular Biology and Evolution. 17(9):1286-1293.
Or what about gene duplication?
Molecular genetics of ecological diversification: duplication and rapid evolution of toxin genes of the venomous gastropod Conus. by Duda TF Jr, Palumbi SR.
quote:
Predatory snails in the marine gastropod genus Conus stun prey by injecting a complex mixture of peptide neurotoxins. These conotoxins are associated with trophic diversification and block a diverse array of ion channels and neuronal receptors in prey species, but the evolutionary genesis of this functional diversity is unknown. Here we show that conotoxins with little amino acid similarity are in fact products of recently diverged loci that are rapidly evolving by strong positive selection in the vermivorous cone, Conus abbreviatus, and that the rate of conotoxin evolution is higher than that of most other known proteins. Gene duplication and diversifying selection result in the formation of functionally variable conotoxins that are linked to ecological diversification and evolutionary success of this genus.
Strong positive selection...
Seems that the data are out there for anyone to look at and they do not seem to support most of Peter's claims.
Yes, many of the examples show non-random mutations, but none of them seem to contradict Neo-Darwinian concepts of randomness, in fact as I have shown neo-darwinian randomness does not refer to location etc but rather to the effect.
Ok one more reference to cone snails
quote:
Espiritu DJ, Watkins M, Dia-Monje V, Cartier GE, Cruz LJ, Olivera BM. (2001) Venomous cone snails: molecular phylogeny and the generation of toxin diversity. Toxicon, 39: 1899-1916.
Abstract: In order to investigate the generation of conotoxin diversity, delta-conotoxin sequences from nine Conus species were analyzed in the context of their phylogeny. Using a standard molecular marker, mitochondrial 16S RNA, we determined that the delta-conotoxins were derived from three distinct species clades based on the phylogenetic reconstruction of a large set (>80) of Conus species and other toxoglossate molluscs. Four different mechanisms appear to have contributed to the diversity of the delta-conotoxins analyzed: (1) Speciation: delta-Conotoxins in different species diverge from each other (the prepro regions of orthologous genes somewhat more slowly than the reference rRNA rate, the mature toxin regions significantly faster). (2) Duplication: Intraspecific delta-conotoxin divergence is initiated by gene duplication events, some of which may have predated the species itself. (3) Recombination: A novel delta-conotoxin may arise through recombination of two parental delta-contoxin genes. (4) 'Focal hypermutation': This sudden, almost saltatory change in sequence is always restricted to the mature toxin region.The first three have been recognized previously as mechanisms important for the evolution of gene families in other phylogenetic systems; the last is a remarkable, mechanistically unexplained and specialized feature of Conus peptide diversification.
So there is hypermutation and while interesting it hardly suggests that the mutations are non random in the Neo-Darwinian meaning of the term.
Constant and hypervariable regions in conotoxin propeptides. Woodward SR, Cruz LJ, Olivera BM, Hillyard DR.
Speciation of cone snails and interspecific hyperdivergence of their venom peptides. Potential evolutionary significance of introns.
Olivera BM, Walker C, Cartier GE, Hooper D, Santos AD, Schoenfeld R, Shetty R, Watkins M, Bandyopadhyay P, Hillyard DR.
And perhaps the most relevant one
Mechanisms for Evolving Hypervariability: The Case of Conopeptides by Silvestro G. Conticello et al.
"Conopeptide Gene Trees Show Evidence of Recent
Diversifying Selection"
[ 24. April 2003, 01:35: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Mesk
Member
Member # 630
|
posted 24. April 2003 02:09
Alpha-actinin-3 is not redundant.
I was quite excited to see a mention of alpha-actinin-3 on ISCID, because this protein has actually been my sole research focus for the last year-and-a-half and is the topic of ongoing research in my lab. The only two major recent publications on this protein (North et al. 1999 and Mills et al. 2001) have come from the lab I work in, so I can discuss it with a fair bit of authority.
I see that someone has already cited one of my conference abstracts on our research into the frequency of alpha-actinin-3 deficiency in athletes. Since that abstract our research has progressed nicely, and we have now submitted a manuscript to American Journal of Human Genetics for publication. Since our data are still technically unpublished I can't give a lot of detail, but I can summarise our basic approach and major findings.
First some basic background: the sarcomeric alpha-actinins (2 and 3) are major structural components of the contractile apparatus in skeletal muscle. The alpha-actinin-3 gene, ACTN3, has two alleles in humans: the R allele, which expresses functional protein, and the X allele, which contains a premature stop codon that prevents the expression of protein. Humans which have two copies of the X allele (that is, have an XX genotype) are totally deficient in alpha-actinin-3, and presumably can survive because of compensation by the related protein alpha-actinin-2. Humans who have RR or RX genotypes do express detectable alpha-actinin-3 (although RX subjects express about half as much). In our study we looked at the frequency of the three different genotypes in a large group of controls and in a group of elite athletes.
Our results were quite unambiguous: alpha-actinin-3 is not functionally redundant in humans. The frequency of the XX genotype (which causes deficiency of alpha-actinin-3) was significantly lower in sprint/power athletes than in controls, indicating that the presence of alpha-actinin-3 is required for optimal performance in sprint/power athletes. Thus alpha-actinin-3 clearly does have functions in humans, and its absence has a deleterious effect on muscle function in high-speed activities. I'm currently attempting to work out the precise mechanisms underlying these effects through the generation of a Actn3 knockout mouse model.
We also discovered hints as to precisely how the X allele - which appears to be disadvantageous for sprint performance - reached its current high frequency in the human population. It turns out that, in contrast to sprint athletes, endurance athletes have a higher frequency of the XX genotype than controls. In other words, it appears that the absence of alpha-actinin-3 somehow provides an advantage for long-distance performance. Thus both alleles of the ACTN3 gene provide advantages, but under different conditions - the R allele benefits sprint performance, while the X allele benefits endurance performance. Exactly how this works will hopefully be partly elucidated using results from the Actn3 knockout mouse.
One other study I'm doing at the moment involves analysing the region surrounding the R/X polymorphism for other genetic variations. The precise pattern of genetic variation in this region will give us information about the history of the two alleles, and provide clues about the sorts of evolutionary forces which have acted on them during that history. I do have some very preliminary data which have got us quite excited, but I won't comment on that until we're ready for publication.
Anyway, in summary - the alpha-actinins do not provide support for Peter's ideas, since alpha-actinin-3 is not functionally redundant and the high frequency of its deficiency in humans can be readily explained by balancing selection acting on both the R and X alleles. This protein is not a problem for evolutionary theory.
Mesk.
[ 24. April 2003, 02:10: Message edited by: Mesk ]
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 02:42
Hi rex,
Thanks for your extensive analysis. Not by head I presume but by using a preexisting computer program?
First of all how does the program exclude a possible common mechanism that introduces the mutations on the same spot? I have given the definition of NRM1 and NRM2. How did you exclude such mutations?
And as pointed out, hotspots are not random, since they can be predicted: they are introduced over and over on the same spot. That's why they are called hotspots. It implies a mechanism. Whether or not protein/RNA mediated is not relevant. It may even be due to the DNA sequences (hairpin formation, chromatin modification, etcetera). The hotspots only become evident after comparing several subspecies.
And regarding the 1g5 gene you cannot apply selection here since the authors show that the complete region is in accord with neutral evolution. So, there is a mechanism underlying the mutaions that introduce them on the same spot. I don't see another solution. Probably in response to a parasite in the environment?
You say that it "won't dramatically alter conclusions about relatedness" and that "hotspots may mess up the accuracy of using base pair changes as molecular clocks, but it tends not to have a huge effect on the topology of phylogenetic trees since hotspots are random, just with a different frequency."
Actually I demonstrated that they are NOT random with respect to position and often nucleotide. Caporale's book also clearly demonstrates that a lot of mutations move away from randomness. The issue is that evo-biologists have ASSUMED that they are random. But now we are starting look at genomes in more detail we realise that they are NOT random. That's why I showed the examples that clearly demonstrate that mutations are not random.
In my opinion the experiments to examine the real character of randomness of mutations have never been carried out properly. For instance the Delbruck-Luria experiments in 1943 (!) that suppose to have demonstrated the random character of mutains did NOT demonstrate that at all. All they demonstrated was that the population of bacteria they studied already harboured preexisting information that gave them resistence to the bacteriophage. They ASSUMED it was due to a random mutation, but I see no reason why to assume it was random.
I really find it amazing that I was able to find the nonrandom character of mutations as soon as 10 or more sequences are compared within the species (intraspecies) and that this is reflected between species (interspecies comparison). You may have the idea that the ZFY region is in agreement with common descent according to your program, but according to my program it is clearly not.
Also the origin of the presented sequences is not dubious: I've obtained them from Dr Page. You must know him: Scott, the primatologist, present on all evo-vs-creo debates.
quote:
"Now, we just have one last thing to check: does this make sense at all given what we know about the histories of the organisms?"
History of the organisms? I think you are missing a point here. This examples only says something about DNA sequences and how they are able to change according to rules. The rules are determined by NRM1. Or you have to exclude NRM1.
As far as I am concerned you didn't do that.
It goes like this: Similar organisms --> same original DNA sequences + same biochemistry --> same rules --> same mutations--> mutations line up--> common descent? NO. Common origin? Yes.
Maybe we could also have a look at the corresponding region on the X chromosme (the ZFX gene) and redo your analyssis. I bet it will demonstrate even stranger things: no change at all for 25 million years, not even on neutral (third codon silent) positions. Selection or absence of a common mechanism? I know what to choose. I go for GUToB.
Mfu and Cae--two types of similar monkey; looks good.
Hsy and Hag--two gibbons; good.
Pt2 and Ggo--chimp and gorilla; good.
Pt1 and Str--chimp and tamarin (?!?!?!) Weird!
Gxx and Ppy--chimp/gorilla and orang; good.
Mxx and Sxx--monkeys and tamarin/chimp; hard to evaluate with chimp (probably good with tamarin).
AncB and Hxx--chimp/gorilla/orang and gibbons. Nicely clustering apes.
AncC vs. Lca--lemurs nicely separated from monkeys/apes
What you demonstrate with the clustering is that it could be a tool to identify the original created multipurpose genomes. And I don't find it surprising that 'strange things' happen to molecular phylogenetic analysis as soon another second (third etcetera) subspecies is introduced.
quote: "The only thing that is bizarre is the supposedly second chimp sample of Ptr1."
In my previous mail I have pointed out what my points of concern were regarding the ZFY and they pertained not only the chimp. I am looking forward to seeing more sequences in distinct subpopulation and whether the mutations line up (as for the 1g5 gene). Maybe you have additional sequences of chimp ZFY?
Even if the gene has been duplicated the chimp still shares the mutations with lemur and tamarin, and still points in the direction of NRM.
quote:
"Personally, I doubt Pty1 is a chimp at all; somewhere along the line, it probably got mislabeled (or was deliberately mislabeled by the list-compiler in order to check whether anyone was actually looking at the data properly)."
As mentioned, I've obtained the sequences from Dr Page. You can say a lot about Page, but he is not a cherry picker, or deliberately mislabeling.
quote:
"common descent seems to be well-verified and the standard theory seems in excellent agreement with this data set."
That's exacly it: Well verified. It is also very easy to falsify too. So, I don't see an proevolutionary argument here.
Best wishes,
PB
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 02:58
Hi Mesk,
I think we must have met in the Powerhouse museum in Sydney last year. If I remember we even spoke about your findings in athlets. What I recall from your poster is that in the population of studies athletes there is also a subpopulation of sprinters that have the double negative genotype (XX) for the ACTN3 gene. This demonstates the redundancy in this subgroup I would say.
best wishes,
Peter
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. April 2003 05:57
PB, I did the phylogenies by hand. The only program involved was a text editor (lots of cut and paste). I more-or-less explained how I did it as I went through.
Pty1 is mislabeled in GenBank as a ZFY. In the text of the paper (Kim & Takenaka, Mol. & Cells 10(5):512-18), it is said to be ZFX. This is easily verified by checking against the gorilla ZFX (accession numbers AB041909 and AB041907). I have notified GenBank and the authors. You may wish to inform Dr. Page.
code:
Top: "Pty1"
Bottom: gorilla
1 gaccagcaag gcagagaagg ccattgaatg cgatgagtgt gggaagcatt tctctcatgc
1 gaccagcaag gcagagaagg ccattgaatg cgatgagtgt gggaagcatt tctctcatgc
61 aggggctttg tttactcaca aaatggtgca taaggaaaaa ggagccaaca aaatgcacaa
61 aggggctttg tttactcaca aaatggtgca taaggaaaaa ggagccaaca aaatgcacaa
121 gtgtaaattc tgtgaatacg agacagctga acaagggtta ttgaatcgcc acctcttggc
121 gtgtaaattc tgtgaatacg agacagctga acaagggtta ttgaatcgcc acctcttggc
181 ggtccacagc aagaactttc ctcatatttg tgtggagtgt ggtaagggtt ttcgtcaccc
181 ggtccacagc aagaactttc ctcatatttg tgtggagtgt ggtaagggtt ttcgtcaccc
241 gtcagagctc aaaaagcaca tgagaatcca tactggggag aagccgtacc aatgccagta
241 gtcagagctc aaaaagcaca tgagaatcca tactggggag aagccgtacc aatgccagta
301 ctgcgaatat aggtctgcag actcttctaa cttgaaaacg catgtcaaaa ctaagcatag
301 ctgcgaatat aggtctgcag actcttctaa cttgaaaacg catgtcaaaa ctaagcatag
361 taaagagatg ccattcaagt gtgacatttg tcttctg
361 taaagagatg ccattcaagt gtgacatttg tcttctg
Hopefully you will revise your statements accordingly.
Note that the standard tree-building model predicted that this data point was in error, and in fact, it was. Your model failed to make this prediction.
Now, on to details.
I didn't exclude a mechanism that introduces mutations on the same spot. Rather, I assumed that the best tree would be the one that minimized the number of mutations--and found that tree using only half the data. The tree was basically confirmed by the other half of the data.
Are you trying to say that there are mutations that occur simultaneously in pairs of unrelated (but similar) species, in a coordinated fashion, separated by a couple hundred base pairs? That is what it would take for common mechanisms to give the same trees in both halves of the data. It is exactly the confirmation by the other half of the data that rules out NRM messing up our phylogeny.
The Luria-Delbruck experiment confirmed random mutation because of the distribution of frequencies of survivors. I suggest you read the paper, if you haven't already. It's a striking confirmation of exactly the prediction you'd get from random mutation (with respect to future/current environmental change). It makes no prediction about where the mutation is, hot spots, etc. etc.; it just shows that the mutation is not predictive of the environment or responsive to the environment, and that the rescuing mutation is approximately equally likely to happen to any organism at any generation. This is exactly what is meant by "random mutation" in evolution.
Hot spots are, by definition, spots. Thus, they are not "random" with respect to position; they occur at defined positions. And they needn't be random with respect to nucleotide either. This isn't exactly a surprise to those of us familiar with mutations. It also isn't the least bit troubling to a Darwinian view of evolution.
Of course, you can postulate millions or billions of mechanisms for every base pair change that is shared between organisms that you don't want to be related, but molecularly appear to be. However, this (1) has no predictive power, (2) fails to account for the predictive power of the standard method, and (3) is a case of rampant overfitting to data, as measured by number of free variables to number of observations. This isn't to say that mutations don't occur faster on, say, ZFY than ZFX. Being on the Y chromosome is rather a weird thing for a gene, though--the paper seems to indicate that ZFY is a candidate for hypermutation in exons. Fine--much of ZFY is a "hot spot". This doesn't really do much to the Darwinian hypothesis or common descent. As noted above, we can even use that to find errors in our data.
Added in edit: if you do inform Dr. Page, please give me credit. Building the trees, finding the error, and verifying it wasn't entirely trivial.
[ 24. April 2003, 06:02: Message edited by: Rex Kerr ]
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 06:25
Hi Rex,
I doubt whether things get more transparant now, since now it seems that the ZFX region in Gorilla has the same positional mutations as observed in the ZFY region for lemur and tamarin. Even more peculiar. And according to Kim's paper the ZFX gene is completely stable in primates.
Best wishes,
PB
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 07:39
Hi Rex,
What I do not understand is where does your model predict that mutations take place at the same spot in the sequences?
Although you claim that Lemur and Tamarin are way of the rest of these primates, the mutations are in exactly the same spot of the gene. My point is that over and over the same nucleotide is involved. For instance TA (chimp) becomes TG (Orangutan) becomes CG (gibbon) becomes CT (lemur/tamarin). And over and over exactly the same positions are involved. The same holds for all other shared mutations. Whether or not you are able to make a tree with these data is not that relevant. What is relevant is there an undelying mechanism that introduces the mutations over and over on exactly the same spot. Why is that? That's NRM1.
Quote:
"Are you trying to say that there are mutations that occur simultaneously in pairs of unrelated (but similar) species, in a coordinated fashion, separated by a couple hundred base pairs?"
I am not only trying to say this, but the data demonstrate that over and over the same positions mutate.
In addition, I am trying to convey that NRM does not mess up common descent but it is the origin of the illusion of common descent:
Similar organism --> same original genes-->same molecular mechanism--> same shared mutations.
That explains why the positions are always the same but the nucleotide is variable. Your theory would rather predict that both position and nucleotide are random throughout the sequences. Furthermore, GUToB does not exclude random mutations. But it acknowledges that there are also NRM. And these NRM give the illusion of common descent.
quote:
"The Luria-Delbruck experiment confirmed random mutation because of the distribution of frequencies of survivors."
No, you cannot say that. It showed that someting random had happened. But it most probably is random loss of the DNA elements that give resistance to the phage. Random loss is much easier to conceive than the appearance of new DNA elemenst that induce resistance. It should be noted that the initial organisms they worked with had not been grown in the presence of the phage and so there was no selective constraint to keep these DNA elemenst in the genome. They didn't exclude this possibility.
You call it a rescueing mutations, but I simply claim that the DNA elements involved are part of the multipurpose genome (MPG) of bacteria (and the bacterial MPG is rather big since they be regarded as one or only a few MPGs).
You also think that hotspots aren't the least bit troubling to a Darwinian view of evolution. As pointed out (again) the mutations you analysed for the ZFY region are non-random with respect to position and random with respect to nucleotide. The latter observation is very important since it tells me that a mechanism does the trick. What else could explain that over and over the same positions are involved? Let me ask you this: why would the same positions be involved all the time?
What I don't understand is where you get the idea that I don't want organism to be related. Of course they are related by DNA sequences and by biochemical mechanism. What I fail to see is that such mechanisms proof common descent. And as mentioned there are a lot of DNA sequences not showing common descents, it simply depends on the DNA regions one studies.
Concerning you points concenring GUToB:
(1) has no predictive power,
Tomorrow I will send the rules of GUToB, what it predicts and where it is superior to the standard theory.
How was it again for redundancies and the predictive power of the standard theory. Or the explanation for the swimreflex in conjunction with the gag reflex--> Natural selection would be my guess.
(2) fails to account for the predictive power of the standard method,
Depend on what you call predictive power. If you mean verification. Okay. Let's have look at ubiquinine, histons, or the ZFX gene. I don't see any predictive power for these DNA elements.
(3) is a case of rampant overfitting to data, as measured by number of free variables to number of observations. This isn't to say that mutations don't occur faster on, say, ZFY than ZFX.
Not only faster or slower, they are absent for 25 million years also on neutral position. What kind of evolution is that?
Quote:
"Added in edit: if you do inform Dr. Page, please give me credit. Building the trees, finding the error, and verifying it wasn't entirely trivial."
I am afaid that Dr Page doesn't listen to me.
Best wishes,
PB
IP: Logged
|
|
Pim van Meurs
Member
Member # 541
|
posted 24. April 2003 13:17
Peter you state that: And according to Kim's paper the ZFX gene is completely stable in primates. And also that they are on 'neutral positions'.
If I remember correctly Kim's paper looked at a partial exon and only a small number of bases so how did you reach your conclusion about completely stable and 'neutral positions'?
From the abstract
quote:
We have sequenced the partial exon of the zinc finger genes (ZFX and ZFY) in 5 hominoids, 2 Old World monkeys, 1 New World monkey, and 1 prosimian. Among these primate species, the percentage similarities of the nucleotide sequence of the ZFX gene were 96-100% and 91.2-99.7% for the ZFY gene. Of 397 sites in the ZFX and ZFY gene sequences, 20 for ZFX gene and 42 for ZFY gene were found to be variable. Substitution causes 1 amino acid change in ZFX, and 5 in ZFY, among 132 amino acids. The numbers of synonymous substitutions per site (Ks) between human and the chimpanzee, gorilla and orangutan for ZFY gene were 0.026, 0.033, and 0.085, respectively. In contrast, the Ks value between human and hominoid primates for the ZFX gene was 0.008 for each comparison. ***Comparison of the ZFX and ZFY genes revealed that the synonymous substitution levels were higher in hominoids than in other primates. The rates of synonymous substitution per site per year were higher in the ZFY exon than in the SRY exon, and higher in the ZFY exon than in the ZFY intron, in hominoid primates***
There is another paper which may be of interest to you here as well
Is selection responsible for the low level of variation in the last intron of the ZFY locus?
Jaruzelska J, Zietkiewicz E, Labuda D.
Mol Biol Evol 1999 Nov;16(11):1633-40
quote:
DNA variability was investigated in the last intron of the Y-chromosome-specific zinc finger gene, ZFY, and its X homolog on Xp21.3, ZFX. No polymorphisms were found in the 676-bp ZFY segment in a sample of 205 world-wide-distributed Y chromosomes, other than a solitary nucleotide variant in one individual (nucleotide diversity pi = 0.0014%). In contrast, 10 segregating sites (pi = 0.082%) were identified within 1,089 bp of the ZFX sequence in a sample of 336 X chromosomes. Four of these polymorphisms, which contributed most of the diversity, were located within an Alu insert disrupting the ZFY-ZFX homology (pi Alu = 0.24%). The diversity in the homologous portion of the ZFX intron, although higher than that in ZFY, was lower than that found in genomic segments believed to evolve neutrally; interspecies divergence in both segments was also reduced. Although this suggests that the evolution of both ZFY and ZFX homologs may not be entirely neutral, both Tajima and HKA tests did not reject neutrality. The lack of statistical significance may be attributed to a lack of power in these tests (the low divergence and variability values reduce the power of the HKA and Tajima tests, respectively); furthermore, Homo sapiens has recently undergone a rapid population growth, and selection is more difficult to detect in an expanding population. ***Therefore, the failure to reject neutrality does not necessarily indicate the absence of selection.*** In this context, the phylogenetic argument was given more weight in out interpretations. The high level of sequence identity in ZFY and ZFX segments, in spite of their separation 80-130 MYA, reflects a lower mutation rate as compared with other segments believed to undergo unconstrained evolution. Thus, the possibility of weak selection contributing to the low level of nucleotide diversity in the last ZFY intron cannot be excluded and should be kept in mind in the population genetics studies based on Y chromosome variability.
Other papers of interest may include
Mol Biol Evol 1998 Feb;15(2):138-42 Sequence variation in ZFX introns in human populations.
Huang W, Fu YX, Chang BH, Gu X, Jorde LB, Li WH.
Mol Biol Evol 1993 Mar;10(2):271-81 Evolution of the Zfx and Zfy genes: rates and interdependence between the genes. Pamilo P, Bianchi NO.
quote:
A phylogenetic analysis of sex-chromosomal zinc-finger genes (Zfx and Zfy) indicates that the genes have not evolved completely independently since their initial separation. The sequence similarities suggest gene conversion in the last exon between the duplicated Y-chromosomal genes Zfy-1 and Zfy-2 in the mouse. There are also indications of conversion (or recombination) between the X- and Y-chromosomal genes in the crab-eating fox and in the mouse. The method for estimating synonymous and nonsynonymous substitutions is modified by incorporating the substitutions in the twofold-degenerate sites in a novel way. The estimates of synonymous substitutions support the generation-time hypothesis in that the obtained rates are higher in mice (by a factor of 4.7) than in humans and higher in the Y-chromosomal genes (by a factor of 1.9) than in the X-chromosomal genes.
J Mol Evol 1994 Dec;39(6):569-78 Contrasting rates of nucleotide substitution in the X-linked and Y-linked zinc finger genes. Shimmin LC, Chang BH, Li WH.
quote:
We have sequenced the entire exon (approximately 1.180 bp) encoding the zinc finger domain of the X-linked and Y-linked zinc finger genes (ZFX and ZFY, respectively) in the orangutan, the baboon, the squirrel monkey, and the rat; a total of 9,442 bp were sequenced. The ratio of the rates of synonymous substitution in the ZFY and ZFX genes is estimated to be 2.1 in primates. This is close to the ratio of 2.3 estimated from primate ZFY and ZFX intron sequences and supports the view that the male-to-female ratio of mutation rate in humans in considerably higher than 1 but not extremely large. The ratio of synonymous substitution rates in ZFY and ZFX is estimated to be 1.3 in the rat lineage but 4.2 in the mouse lineage. The former is close to the estimate (1.4) from introns. The much higher ratio in the mouse lineage (not statistically significant) might have arisen from relaxation of selective constraints. The synonymous divergence between mouse and rat ZFX is considerably lower than that between mouse and rat autosomal genes, agreeing with previous observations and providing some evidence for stronger selective constraints on synonymous changes in X-linked genes than in autosomal genes. At the protein level ZFX has been highly conserved in all placental mammals studied while ZFY has been well conserved in primates and foxes but has evolved rapidly in mice and rats, possibly due to relaxation of functional constraints as a result of the development of X-inactivation of ZFX in rodents. The long persistence of the ZFY-ZFX gene pair in mammals provides some insight into the process of degeneration of Y-linked genes.
Where did you get the idea that the ZFX gene is "completely stable"?
As far as non-randomness you state once again "You also think that hotspots aren't the least bit troubling to a Darwinian view of evolution. As pointed out (again) the mutations you analyse"
And as I have shown neo-Darwinian theory accepts the non-randomness of mutations with respect to location so indeed they are not the least troubling to a Darwinian view of evolution.
It's that simple really, your appeal to non-randomness has no relevance to how the term is used in neo-Darwinian theory when it states that mutations are random wrt the environment they are in.
[ 24. April 2003, 13:27: Message edited by: Pim van Meurs ]
IP: Logged
|
|
Argon
Member
Member # 276
|
posted 24. April 2003 14:53
If we're worried that identical point mutations may arising independently in separate lines and cloud determinations of descent why not look at pseudogenes, chromosomal mapping and insertion sequences instead? Then check for congruence between gene trees and the trees of the larger structures. If they are largely the same....
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. April 2003 19:02
PB, I don't think you have yet managed to grasp the essential features of molecular phylogenetic analysis, since you haven't managed to address any of the points I raised there.
Chimp doesn't become orangutan, and that doesn't become gibbon. That's nonsense in an evolutionary model. If you are going to critique or even understand competing theories (i.e. the standard, well-verified, well-accepted one), you at least need to be accurate in your statements. Try again: what must have happened in an evolutionary process to observe those nucleotide patterns? You may refer to my trees, since one (hypothetical) answer is there.
In fact, the same positions are not involved over and over again--not according to the standard model. There was a bias of 10 double-hits instead of the 5 expected by chance--see the bottom of my last tree. That's a small increase in positions being involved "over and over again". If you remove the bad Pty1 data point, you're down to 8 vs. the expected 5. That's a very modest increase.
You ask why the same spots might be involved over and over again. It's hard to say, but keep in mind that a factor of 2 corresponds to a tiny amount of energy--perhaps 1/20th of what is available to a cell with a single ATP hydrolysis. What kind of factors could produce that energetic difference? Well--tension on the DNA due to histone binding making it less/more susceptible to UV; alterations in the dwell time of DNA polymerase during replication or transcription of that sequence; binding affinity of mismatch repair machinery to that particular sequence; and so on. Which are actually involved? I have no idea. Probably mostly factors that I haven't thought of. Factors that are reproducible but unspecified; accidents of molecular energetics. There are a lot of accidental energetics/accessbility/etc..
You claim, over and over, that you can get the illusion of common descent from NRM. Please demonstrate this. There are two ways I can think of to do this:
(1) Using an abstract model, show that tree-building produces an illusion of common descent across reasonable sequence lengths when none actually exists.
(2) Show where in the ZFY tree there are areas where common mutations are clearly the cause of a change at a certain position, and that this is being misinterpreted by the tree-generating algorithm as common descent.
Failing something like that, you're just stating opinion. You keep pointing to individual base pair changes, but you have not yet examined this in the context of tree-building. (Nor does your TA/TG/CG/CT example obviously give an incorrrect phylogeny.)
I can't say much about ZFX as I don't have the data for it, and I don't care to dig it all up myself. Maybe there is something troubling there; maybe not. It's up to you to convincingly point it out.
However, there are changes to ZFX. They are just less common. You may wish to refer to the paper for the rates. (Or see Pim's posting.)
With respect to the Luria-Delbruck experiment, random loss produces a completely different pattern than random gain when the net result is that almost everything is lost. Further, in random loss, the initial bacteria should have appeared to be resistant to the phage, which I suspect that they checked (although I don't have the paper in front of me to verify).
The evolutionary model predicts that highly selected-for genes--those critical to the function of life--will vary slowly due to mutations being selected against. This is observed in the critical genes encoding ubiqutin, histones, and ZFX.
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 20:05
Hi Argon,
I already did that. The GLO pseudogene that is usually presented as the ultimate evidence for common descent (since it has a indel-mutation in exon X on the same spot in all higher primates and is supposedly the cause for the inability of synthesising vitamin C).
And guess what. Several positions in the gene show a similar NRM pattern as observed for all them other genes I checked. Thus, NRM in the GLO gene can therefore not be excluded. Moreover, Lynn Ripley demonstrated that indel mutations can be introduced over and over on the same spot due to imperfect hairpins in DNA sequences. So, I don't see a reason to present GLO as evidence for common descent. It could as well be a common mechanism. I even consider that more plausible.
(For a detailed analysis of the GLO pseudogene don't hesitate to ask.)
Best wishes,
Peter
IP: Logged
|
|
peter borger
Member
Member # 722
|
posted 24. April 2003 20:55
Dear Rex,
I notice that you try to introduce the first strawman. I do NOT say that "Chimp become orangutan, and that orangutan become gibbon".
I say according to my molecular biological insights there could be a mechanism at work that introduces over and over mutaions on the same spot and that will give the illusion of common descent. I illustrated that with the two nucleotides that keep changing.
quote:
"Try again: what must have happened in an evolutionary process to observe those nucleotide patterns? You may refer to my trees, since one (hypothetical) answer is there."
I think that if your model is okay the DNA sequences of a putative ancestor of all the presented primates was the same. Next, they start to diverge through random accumulation and selection of mutations (More probably is through chromosomal rearrangement: a non gradual-event). This randomness should be reflected in the genome, but as shown it is not.
On the other hand the GUToB says that 1) the organisms have initially the same DNA elements comprising their multipurpose genome, 2) speciation is immediate (and thus non gradual) through chromosomal rearrangements and 3) due to the same biochemistry and DNA sequences the muations line up and give the illusion of common descent. It explains all biological phenomena. Including genetic redundancies that appear to be present in the multippurpose genome without selective constraints.
quote:
"If you remove the bad Pty1 data point, you're down to 8 vs. the expected 5. That's a very modest increase."
What happened to the ZFX sequence in gorilla that it has exacly the same positional mutations (=NRM) as the ZFY gene of Lemur and tamarin? It seems to me as if the DNA sequences determined these mutations. What's the standard theory explanation?
Ans you provide some mechanism that might be well responsible for NRM1:
1) histone binding making it less/more susceptible to UV;
2) alterations in the dwell time of DNA polymerase during replication or transcription of that sequence;
3) binding affinity of mismatch repair machinery to that particular sequence; and so on.
quote:
You claim, over and over, that you can get the illusion of common descent from NRM. Please demonstrate this.
I demonstrated the NRM observed in living organisms. For instance the 1g5 gene in Drosophila. I have explained the NRM here in detail. The region is even a neutral one according to the authors so it definitely is evidence for a mechanism that is able to introduce mutations over and over on the same spot. Now you ask me to build an abstract model, while I demonstrated already emperical evidence of such a model. I will think about it, but it really puzzles me why. I already presented the real stuff.
Furthermore I do not claim that tree-building produces the illusion of common descent, I claim that similar mechanism operable in similar organisms (=same DNA sequences + same biochemistry) produces the illusion of common descent. And the NRM in ZFX in gorilla and ZFY genes of lemur and tamarin advocate this view.
quote:
"However, there are changes to ZFX. They are just less common."
A stable coding DNA element that doesn't demonstrate mutations on silent positions for 25 million years is NOT evolving. Histons are not evolving. Ubiquitins are not evolving, etcetera. So if they are not evolving when did they evolve?
If there has ever been a proces of evolution than it has finished now.
quote:
"With respect to the Luria-Delbruck experiment, random loss produces a completely different pattern than random gain when the net result is that almost everything is lost."
With NRM I did NOT mean as a response to the environment. That could be excluded in the experiments since other wise there wouldn't have been any survivers. However, the DNA elements to yield the resistance were however already present in the genome and had to be activated. Probably similar to the Rosenberg finds: induction of an alternative errorprone DNA polymerase. And since such polymerase are redundant when cultured without constraints they will get easily lost. So, it depends on several factors: loss of the error prone DNA polymerase and loss of the DNA elements that conveys the resistance. As mentioned, a bacteria can spawn billions of progeny in 24 hours and induction of the polymerase will likely yield an adaptive phenotype. As if it dropped out of the sky! But that's also an illusion: the DNA sequences were already present.
Best wishes,
Peter
IP: Logged
|
|
|
Printer-friendly view of this topic