|
Author
|
Topic: Cytosine Deamination from Both Sides
|
yersinia
Member
Member # 324
|
posted 18. April 2003 08:45
Hi Mike,
What would you say about the potential objection that any kind of designing that could be done via front-loading with mutational biases could be done faster, better, cheaper, and with a few orders of magnitude more certainty in the results, simply by directly designing whatever you're trying to design (whatever that is, you have not clarified).
This very complex indirect design method you propose seems to me like proposing the equivalent of typing with your elbows: it might work, kinda, with an amount of effort ridiculously disproportional to the very uncertain payoff...but why would anyone bother, when you can type with your fingers?
Is anything gained by proposing a spectacularly crude and inefficient method of design as an alternative hypothesis to that of Poole et al.? It may technically be true that one can think up some reason an engineer might prefer cytosine -- but the engineer ends up being so thoroughly bizarre that Poole et al.'s point is proved.
yersinia
PS: I also have severe doubts about how much "information" can possibly be encoded in a general mutational bias -- this is not like a zipfile here. I'm having trouble finding the words, but if you think about what natural selection or direct design can do -- pick the one good mutational change out of a sea of harmful ones and fix it in the population -- in comparison this version of front-loading seems to lack the "specificity" to do much at all. The biggest effect would be in neutral changes (a mutational bias would increase the chance of fixation in the population from very very very low to very very low for any given mutation) -- but these are not likely the important changes anyhow, being neutral.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 18. April 2003 10:07
Nic,
Front-loading, by definition, is about designing the future through the present. It is about imposing some kind of constraint on evolution, or simply put, using evolution to carry out design objectives. The thesis stems from taking the design of the original life forms seriously, as the objectives behind these designs might extend beyond the proximal activity of the original life forms. As an investigator, I'm simply feeling out the plausibility of such speculations.
Your objection is not against any particular formulation of front-loading, but against the concept itself, as you seem to be arguing that any rational designer would design things directly rather than indirectly. But this objection is built upon many assumptions. Remember, I am trying to keep such assumptions to a minimum and thus consider only a human-like intelligence behind the designs. Thus, if we consider a modest extension of Crick and Orgel's thesis of directed panspermia, it's one thing to seed a planet with designed life forms, but yet another thing to return agains and again and directly implement new designs every time they are needed. Taken to its extreme, the logic of your objection would have the designer micro-managing everything.
I also have severe doubts about how much "information" can possibly be encoded in a general mutational bias -- this is not like a zipfile here.
Which is why I have always linked such a bias to carefully chosen original states and thus far propose only a modest unpacking of such designs. An ID thesis does not have to account for every biotic feature to be valid.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 18. April 2003 19:10
Mike, perhaps you missed it in my first post, but the data I cited suggests that C->T substitutions seem to occur less frequently than you would imagine from chance between humans and chimps. (Assuming that human variation is a better measure of chance.) Of course, on the genomic level we can't yet say whether these differences are important for the phenotypic difference between humans and chimps, but the tentative conclusion seems to be, "Nope, C->T wasn't helpful."
Also, while it is an interesting observation that positive feedback can cause an increase in the size of isochores, I am unsure how this relates to any possible design hypothesis. As far as I know, isochores spread equally over both coding and noncoding sequences. If C->T transitions were useful for protein production, you might expect to find protein coding sequences concentrated there.
Early evolutionary questions are notoriously difficult to address. I tend to be quite skeptical of all claims made about early evolution. One can, of course, envision a scenario where genes are silenced until C->T substitutions unlock them, three billion years ago. However, how would one test this hypothesis?
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 20. April 2003 11:45
Rex, again I need to emphasize that I don't propose C->T transitions as the driving force behind genomic evolution. Clearly, there are many factors that come into play. The paper you cited mentions weak directional selection or biased gene conversion in favor of high GC content that could theoretically counteract a predisposition for C->T transitions.
Keep in mind that I also cited a paper that surveyed 2.5 million SNPs that showed the "proportions of substitutions were A/G, 32.77%; C/T, 32.81%; A/C, 8.98%; G/T, 9.06%; A/T, 7.46%; and C/G, 8.92%." (Zhao Z, Boerwinkle E. Neighboring-nucleotide effects on single nucleotide polymorphisms: a study of 2.6 million polymorphisms across the human genome. Genome Res 2002 Nov;12(11):1679-86).
Of course, on the genomic level we can't yet say whether these differences are important for the phenotypic difference between humans and chimps, but the tentative conclusion seems to be, "Nope, C->T wasn't helpful."
The tentative conclusion is not based on any real experimental work. Until we identify the genomic changes (coding or non-coding) responsible for the phenotypic differences, there is no conclusion to be made.
As for the isochores, the interesting feature (IMO) is not whether it is targeted to coding or non-coding sequence, but that it extends into coding sequence. This effect can then be amplified through transcription by the mutational bias it imposes (explained in previous essays). What's more, the positive feedback cycle discussed in that paper is temperature dependent, allowing a designer the possibility of using cytosine deamination to unpack designs coincident with the appearance of homeothermy.
This all goes back to something I wrote in January: Nevertheless, it looks as if I once again need to remind people that my stuff is not only tentative, but represents the very first steps of an investigation. I'm certainly not arguing that all of evolution (or even design through evolution) is carried out through cytosine deamination. The IHE, along with the initial states, are merely some of the design components at play. And the effects of cytosine deamination can be explored through many different potential design considerations, such as its temp-dependence, the role of methylation/CpG dinucleotides, RNA editing, the influence of UV light, the influence of NO, regulation of ung, etc. There is great opportunity for all kinds of research here.
Finally, my hypothesis does not entail the unlocking of silenced genes. The hypothesis entails that the originally designed genes could have buried secondary functional potential that could be untapped through C ->T mutations. That is, you design a protein such that the IHE could tip it into a new functional realm, a function that might be extremely beneficial in the formation of a novel trait.
[ 20. April 2003, 11:46: Message edited by: Mike Gene ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 20. April 2003 11:48
I wanted to extend a thought regarding the relationship between the increasing hydrophicity effect, cytosine deamination, and the genetic code. In my original essay, I noted that the only example where C-T transitions do not increase the hydrophobicity of the amino acid encoded centers around proline (figure 3 from my original essay). I noted, however, that proline would be an important residue to remove if we are to design new secondary structures and the C-T transitions do elicit a remarkable trend toward residues that are usually involved in the formation of secondary structure.
However, one of the things I did overlook in the original essay was the influence of multiple rounds of C-T transitions. If we factor this into the picture, the IHE becomes even more clear.
There are four codons that specify proline and all are C-rich. In fact, 9/12 positions among the proline codons are C. If we envision multiple rounds of deamination (over time), proline changes as follows.
With one substitution, the four proline codons can expand into nine codons - 4 for leucine, four for serine, and one remains proline.
The four codons for serine can become five codons with another C-T transition. Of these five, two become phenylalanine and two become leucine. One remains serine.
The remaining serine codon becomes phenylalanine with the final substitution.
If we go back to the single codon whose coding remained for proline after one round of substitution, a further round gives us serine and leucine. With another substitution, that serine becomes leucine. Thus, the patterns of change we see as mediated by C->T transitions are:
PRO -> LEU
PRO -> SER -> LEU
PRO -> SER -> PHE
In other words, proline is changed to the two most hydrophobic amino acids. And if we return to secondary structure, proline is a strong helix breaker (and a beta strand breaker). Both leu and phe are considered helix and beta strand formers.
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 21. April 2003 01:01
At this point I suppose I'll just have to say that it's an interesting idea. Right now it seems to lack any positive evidence, but perhaps that will change at some point.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 21. April 2003 15:38
Mike wrote:
quote:
Finally, Nelson brings up an interesting hypothesis that is related to something I wrote earlier
With respect to that idea you wrote earlier, I think that RIP genes in Neurospora may be relevant, it may also be an area where one can formulate predictions using that idea.
quote:
Repeat-induced point mutation (RIP) is a process that efficiently detects DNA duplications prior to meiosis in Neurospora crassa and peppers them with G:C to A:T mutations.
DNA methylation associated with repeat-induced point mutation in Neurospora crassa.
Mol Cell Biol 1995 Oct;15(10):5586-97
There is also an interesting article here that may be relevant
[ 21. April 2003, 17:00: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 21. April 2003 22:42
I realize there is no positive evidence for my idea in the sense that there are no documented examples where the C->T transition/IHE have been found to play a crucial role in the evolution of a new function (or better yet, subfunctionalization). Yet we must remember that no one has been looking for such evidence. Or, those who are aware of such data do not have this theoretical backdrop to appreciate its significance (from an ID viewpoint, at least). I do, however, want to extend this analysis in a somewhat different direction because not only is it relevant to the discussions Nelson and Rex had earlier in this thread, but because it appears things get even more interesting from the perspective of FLE. I'm trying to decide whether I should just blurt it out here or write up a more structured essay for TeleoLogic.
And thanks for the heads up about RIP, Nelson. ![[Smile]](smile.gif)
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 22. April 2003 00:17
There also isn't much evidence that any of the effects of C->T substitution are, in general, beneficial to an organism. One could set things up so that they were, but this is true of any substitution. For example, C->T tends to generate stop codons; is this good?
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 23. April 2003 18:59
Mike wrote:
quote:
I'm trying to decide whether I should just blurt it out here or write up a more structured essay for TeleoLogic.
And thanks for the heads up about RIP, Nelson.
Something tells me you are way ahead of me.
Rex,
I think with the paper I referenced, I have shown that a mechanism actually uses C->T mutations in that it seems to have stabilize the genome. Here we have prima facie evidence, at least, that C->T mutations can be put to good use. Also, as has been shown, if you mutate a TTC codon to CCC, a C->T mutation can change CCC to CTC, which actually restores function. Also, that these mutations can be used to generate new and useful sequences is a benefit all of it's own and this use may be found among biological organisms.
[ 23. April 2003, 19:01: Message edited by: Nelson_Alonso ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 23. April 2003 20:38
Rex,
My thesis doesn't predict that C-T transitions would, in general, be beneficial to an organism. The thesis of C-T transitions is tied to the channeling of evolution.
As for generating stop codons, now that's interesting......
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. April 2003 01:24
But you haven't shown evidence that C->T is more interesting for channeling than any other substitution. You can channel with any substitution. How do you intend to show that this one is relevant compared to what you'd expect by chance? Reality doesn't allow all things with equal probability; but you have to specify a target in order to know whether the probability distribution is focused at anything.
(And, at a more sophisticated level, it's nice to be able to distinguish adaptations to arbitrary physical realities, such as C->T being a common substitution, from some sort of foresight, such as channeling evolution somewhere interesting by allowing C->T transitions.)
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 24. April 2003 08:28
Rex,
Have you read my essays on my web page?
IP: Logged
|
|
Rex Kerr
Member
Member # 632
|
posted 24. April 2003 19:10
Yes. You point out some interesting features of cytosine deamination.
But you don't show that those features are more biologically relevant or channelable than other features. You don't examine whether other types of mutation have interesting features.
Unfortunately, for the observations to be meaningful to design/evolution, there has to be a comparison between C->T and something else, at a biologically relevant level.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 25. April 2003 01:50
Rex,
Apparently, we see things differently, as I get the feeling you think I am out to establish/prove design.
Let's start over (then I can get to your points).
Recall the original impetus behind this analysis. Leading scientists asserted, in the peer reviewed literature, that no engineer would have included cytosine in the genetic material. Would you agree that I have successfully called that assertion into question?
It is also often claimed that ID cannot generate any scientific research (i.e., ID is a "goddidit" explanation and nothing more). Would you agree that I have turned back that criticism?
[ 25. April 2003, 01:51: Message edited by: Mike Gene ]
IP: Logged
|
|
|
Printer-friendly view of this topic