|
Author
|
Topic: Response to Jason Rosenhouse
|
charlie d.
Member
Member # 159
|
posted 11. September 2004 00:13
OK, here we go. quote:
I pointed out that the Doring paper [1] provides experimental support for my point. You disagreed, saying that in [1] they required a selection technique to activate the code change. I agreed, and pointed out that there are many models for codon reassignment and that they were testing out one model. Also, that the use of selection hardly disqualifies the method they used anyway. And finally, that no one questions that such reassignments are possible via evolution.
First issue: I did not say that they needed a selection technique to activate the code change. They do not. As soon as the mutant Cys-tRNAs-expressing vectors are introduced in the bacteria, they are active. The ThyA-based selection technique is there for two reasons: first, as a readout for the Cys replacement (it tells the researchers that the constructs expressing the mutant tRNAs work), and second because it allows long-term propagation of the vectors (which otherwise are lost, see figure 2). quote:
First, for Figure 1. For others reading this who are not familiar with the paper, these researchers began by deleting a necessary gene from E. coli (the gene is ThyA which codes for Thymidylate synthase). The interesting thing about this gene, in addition to the fact that it is required, is that codon 146 needs to code for cysteine if the gene is to be active. The researchers supplied an inactive version of the gene which had a different codon at position 146. Actually, they did this four times, testing out four different codons, one for methionine and three for isoleucine.
Obviously, these cells are not going to survive in this state. But the researchers then provided another gene that provides for an alternate DNA code with a single codon change (the gene codes for the cysteine tRNA synthetase, but the translator codon is switched from cysteine to methionine or isoleucine). Specifically, in each case, the codon change was precisely what was needed to activate the necessary gene (ie, make a cysteine go at position 146 in Thymidylate synthase). In other words, they changed the methionine codon and each of the three Isoleucine codons, in turn, to be cysteine. This would activate the necessary gene so the bacteria could survive; however, at the same time, this codon reassignment would be happening in other proteins as well. Overexpression of the DNA-code-changing gene helped to make the change occur as much as possible.
No. The code change is not caused by the Cys-tRNA synthetase gene, but by the mutant Cys-tRNA genes. Expression of those alone is sufficient to suppress the ThyA active site mutation at sufficient rates to make bacteria grow in the absence of thymidine (which otherwise has to be supplied exogenously to ThyA mutant bacteria). However, even in the presence of Cys-tRNAs, the replacement of Cys at the mutant codons is just partial. This is highlighted by the fact that of the 4 mutant tRNAs expressing strains, only one (Cys-tRNA/GAU) allowed growth rates almost as high as normal.
Note that ThyA mutant suppressor activity (i.e., the ability to complement the ThyA mutations) is a very sensitive assay: it only takes a small number (~100) “correct” ThyA molecules (i.e. molecules with a Cysteine at the mutated site) in a cell for it to grow in the absence of thymidine. So what is limiting the the suppression rate from the mutant tRNAs? What's going on is probably that the mutant Cys-tRNAs, although highly expressed, were poorly loaded with Cysteine (other experiments cited in the paper show that Cys-tRNA variants similar to the low-suppressors here are loaded with Cys 2000-fold less efficiently than the wild-type Cys-tRNA). In other words, while the mutant tRNAs were expressed, not nearly a significant fraction of the relevant codons were being replaced by cysteines, the remainder being used by the correct aminoacids Ile or Met, from the wild-type tRNAs which are still present in the cells.
To increase the level of Cys replacement, the authors therefore over-expressed Cys-tRNA synthetase, which is the enzyme that loads Cysteines onto their tRNAs. This leads to mutant Cys-tRNAs being more often “charged” and ready to go, and thus being better competitors with the wild-type tRNAs.
The take-home point here is that in all these strains, Cys replacement at the specific codons (the 3 for Ile and the one Met), are always partial (unlike bona fide code variants). Without over-expressing Cys-tRNA synthetase, this effect is even more marked. quote:
If people have trouble following all that, the bottom line is that the DNA code was altered, though not absolutely 100%, so that 1 out of the 64 codons was now switched to cysteine.
Again, only partially switched. quote:
Across the E. coli's proteome the cysteine amino acid would now be showing up where there was supposed to be a methione or isoleucine, in four separate experiments. This is not exactly a conservative change, as cysteine is, nominally, quite rare. And it is reactive compared to isoleucine. Also, it introduces the possibility of a disulfide bond. Swapping cysteine for isoleucine is not like swapping two similar amino acids.
So what happened? The growth rates of the four DNA code variants were indistinguishable from the control cases at lower temperatures. They had reduced growth rates at high temperatures. And two of the four variants did fine at medium temperatures. Charlie concludes that these results demonstrate that these DNA code variants aren't doing so well. Well, that's true at high temperatures but not at low temperatures, and not in two of the cases at medium temperatures. My point was not that all codon reassignments are workable, nor that codon reassignments are going to always work in all environments. Indeed, what surprised me was that these non conservative reassignments (not 100% but pretty high) worked as well as they did. Six out of the 12 experiments showed the DNA code variants doing quite well. These results support my original point that there are reassignments that are likely to leave the organism working just fine, even though the change was implemented proteome-wide.
I have in fact two points here: first, measuring growth rates in rich medium is not the appropriate way to look at bacterial fitness (which is really what matters in evolution – see below), because these are short-term experiments in optimal growth conditions. They tell you nothing of population dynamics, just of gross cell viability. The second one is that the authors did not look at growth at 42C just on a whim, but because any stress on protein structure caused by the replacement of aminoacid residues is more apparent at that temperature. That is, looking at growth rates at higher temperatures is a more sensitive way to gauge possible deleterious effects of amino acid replacements on protein structure and function.
Indeed, for two of the tRNAs, this effect is so pronounced that it is detectable already at 37C (in the presence of Cys-tRNA sythetase to achieve higher replacement rates). In fact, looking at the top histogram in Figure 1, the negative effect might be already beginning to manifest itself at 30C for Cys-tRNA/GAU (we would need the raw data to actually test the statistical significance of the difference). This would not be surprising, actually, since that is the mutant tRNA that allowed the best replacement rate off the bat (see above). The higher the replacement rate, therefore, the more negative the effect (and, once more, none of these mutants achieves complete replacement as in a bona fide code variant).
These organisms don’t work “just fine” – their proteins are without doubt more unstable, the more so the more misincorporated Cysteines they carry. quote:
Now on to Figure 2. In this experiment, the researchers took the DNA code variant that did the best and grew it for 80 days (or 530 generations). Once a day they had to sample the strain and restart it in a fresh medium to keep things going (serially subculturing). Also, they inserted markers so they could track the inserted genes. Specifically, they provided the gene to change the DNA code and gave it one marker (chloramphenicol-resistance, call this marker CH). And they provided the necessary ThyA gene mutated to work only with the altered DNA code, and gave it another marker (carbenicillin-resistance, call this marker CA).
They also created a control case, where the only difference was that the ThyA gene was not mutated. Therefore, the the gene to change the DNA code would not be needed. Everything else was the same, including the markers.
So what happened? The DNA code variant strain propagated just fine. It grew just as well as the control case, and both markers continued to give signal. In other words, the DNA code change did not go away. They even did an independent check at the end to ensure that the DNA code change had not reverted. It hadn't. The authors concluded that "A single missense mutation thus seems sufficient to perpetuate an ambiguity in the genetic code."
You forgot to say that they applied selection for suppression of the ThyA mutation. That is, the mutant tRNA gene persisted in the ThyA mutant strain when kept under constant positive selection for its (the tRNA’s) expression. quote:
The control case showed that the growth rates were the same between the two strains. Also, in the control case while the CA marker persisted, the CH marker gradually went away. This means that the presence of the gene to change the DNA code gradually decreased. How did this happen? Well, what if at the beginning of the experiment that gene (and its CH marker) failed to make it into every E. coli cell? The population would be mixed, with most E. coli having both inserted genes and a few with only the the ThyA inserted gene (any E. coli that did not have the ThyA gene inserted would have quickly died off). If those few E. coli that lacked the gene to change the DNA code had even an ever so slight higher growth rate, then eventually they would take over and dominate the population. And that is apparently what happened.
Even if that were the case, that would show that cells without the mutant tRNA gene could rapidly outcompete cells bearing the gene (i.e. they had higher fitness). That's the whole point, in evolution. But in fact that is not even the case – all these experiments were started with double resistant (CA/CH) populations. That means that the loss of the tRNA-encoding plasmid was a secondary event in culture. It just makes the case for the lowered fitness of the mutant tRNA-expressing cells even stronger.
Importantly, note that this experiment was run (purposefully, I think) with one of the tRNA mutants (Cys-tRNA/UAU) that had the lowest replacement rates, indeed the one with the lowest effect on growth rates in the previous experiment. This means that even a minimal replacement of Cys causes enough fitness decrease that a mere 500 generations are sufficient for a non-expressing variant to arise spontaneously in the culture and virtually replace (>99%) the original clone. quote:
Charlie suggests that the fact that the code change gene (and CH marker) went away indicates that even a partial code change has enough detrimental effect on fitness that those cells are quickly outcompeted. Well, that may be true, but it may not be. We're talking about a minor change in fitness. Those cells were not "quickly" outcompeted, and the minor code change is not the only difference between the strains. Remember, the code change strain is also carrying an additional gene now compared to the other strain that beat it out.
The only difference is actually the chloramphenicol-resistance encoding gene. These are very well-characterized, very stable genetic elements, used in thousands of experiments, that simply cannot account for this decrease in fitness (if you want to run the control experiment, you can actually do it yourself quite easily in a basic bacteriology lab).
And, from an evolutionary standpoint, this is not at all a minor change in fitness (500 generations to virtual elimination is nothing). It shows that even a very partial code replacement like the one caused by Cys-tRNA/UAU cannot persist without positive selection, and is rapidly eliminated even when starting from a 100% frequency in the population. Remember that in natural conditions the table would be turned - you'd have have the occasional individual code mutant having itself to outcompete an entire population of wild-type organisms. quote:
In any case, even if the minor code change is the cause of the minor fitness reduction, it is still not clear what to make of that. Even known code variants are recognized to have been non adaptive. Remember, the Figure 2 experiment is a highly artificial environment where one of the two strains is likely to become dominant eventually. Consider a one-on-one basketball game. One player is going to win and the other is going to lose. The losing player may, nonetheless, be very good. He may be good enough to be a pro and even to get into the All-Star game. But he isn't as good as the other player. In this artificial environment the minor code change strain eventually goes away, but that doesn't mean it won't survive in natural environments.
This is just wrong. This is nothing like a one-on-one basketball game. I don’t really need to explain why – you are starting with a pure population of a certain kind, then a mutant arises and in less than 500 generations it takes over the place. This is more like a few billions one-on-ones, in which the same guy always wins. quote:
Bottom line is this. The Figure 1 results do not indicate that all the codon reassignments are less fit.
Wrong – figure 1 has nothing to do with fitness (you understand what genetic fitness refers to, right?), just with short-term growth rates in optimal conditions, and even there one can see an effect, even for partial code variants. quote:
The Figure 2 results show that the code change is stable.
Wrong. It shows that it is stable only if it is positively selected. quote:
The fact that the CH marker went away in the control case is not easy to interpret for our purposes and does not prove that the code change variant could not survive, as other code variants must have.
It is in fact very easy to interpret: even very limited amino acid replacements as those caused by Cys-tRNA/UAU cause a significant decrease in fitness. Together with the selection experiments, it shows that a code variant is likely to become stable in a population only when some sort of positive selection takes place, for instance when it can act as a missense suppressor.
I don't know if anyone else has made it through all this - I realize it's quite technical and dry, but there were many errors to correct, and it is important to interpret data for whjat they are, and not for what we wish they were. In fact, I think this is really not too hard of a paper to go through for anyone with minimal knowledge of molecular biology and genetics, as well as some patience with lingo (bacteriologists are notorious for arcane abbreviations). The results are actually very clear.
This is not altogether bad for Cornelius' ideas, though. The current evidence suggests that the code changes tested so far are deleterious, but the question of the extent of tolerance for variants is a fair one. We do know that code variants arise spontaneously all the time, from the work of suppressor mutations in bacteria, so there is no reason not to investigate the question further. I propose Cornelius enlists the help of some decent microbiologist (eg, Minnich), secures some funding from the DI (its lawyers can take a small pay cut, if necessary for ID's sake, can't they?), and start a project using Doennig's approach, on codon substitution for some very conservative replacement (eg, Leu/Ile?). The only tricky part I can see is to identify a selection system which depends on Leu/Ile missense suppression (*); after that, it's all downhill. Try this in bacteria, and maybe in yeast, and publish your results (whatever they show!). That would be at least a step in the right direction.
* Added in edit: Whaddayaknow, that may not even be so difficult. No excuses, then!
[ 11. September 2004, 09:07: Message edited by: charlie d. ]
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 11. September 2004 14:38
Charlie:
Thanks again for your comments on the Doring paper. In trying to keep the jargon down I used the phrase "activate the code change" (as in activate the DNA code change) to describe the process of creating the code change, as distinguished from the question of how well the organism performs once the code change is in place. I now see the reason you perceived this as an error, as you thought I was referring to the initial activation of the code change versus propagating it. That was not the case.
Regarding my one-on-one basketball game, you say it is wrong because it is really many cells vs. one cell. But of course, it is obvious that I used the analogy to represent the strain types, not the cell counts, and that I was talking about a few vs many situation in terms of cell count. (Nonetheless, I do agree that the idea was a stretch.)
Regarding Figure 1 and fitness, you overstate my claim. I said Figure 1 does not indicate that all the codon reassignments are less fit (it does indicate that some are less fit at certain temperatures, but in other cases we cannot draw this conclusion). Of course, this experiment provides limited information, but you are incorrect that it "has nothing to do with fitness." When the results show that a codon reassignment has slower growth rate at certain conditions, then it is a safe bet that it has lower fitness in those conditions. I think what you meant to say is that Figure 1 does not indicate that any codon reassignments, at any temperature, has not lost fitness (in other words, Figure 1 does not reject the possibility that all the codons are less fit). Agreed, but this does not contradict my conclusion.
You also said that the researchers examined cell growth at a higher temperature (42C) "because any stress on protein structure caused by the replacement of aminoacid residues is more apparent at that temperature." Actually, this is not necessarily true. In fact, the researchers stated reason was simply because earlier research had shown that "Misincorporation of amino acids through erroneous codon reading is known to cause heat sensitivity in microorganisms."
I concluded that "The Figure 2 results show that the code change is stable," and you replied "Wrong. It shows that it is stable only if it is positively selected." But, again, since I explained that several times, it should be obvious that that was implied. Obviously, my point was not that there wasn't positive selection, but that the system worked at all.
Then you end with some sarcasm about lawyers (violating your own board-related comment from earlier) and getting ID funding (which makes no sense since ID proponents, if anything, would disagree with my hypothesis).
Having said all that, and while I do agree with you that there are some interesting possible research questions here, I do see your point that the paper does not provide the kind of support I was thinking of. The key point you make, which I misunderstood, was that the cysteine replacements (i.e., the DNA code change) are quite partial. My understanding was that it was pretty efficient, but you make the following point:
quote:
However, even in the presence of Cys-tRNAs, the replacement of Cys at the mutant codons is just partial. This is highlighted by the fact that of the 4 mutant tRNAs expressing strains, only one (Cys-tRNA/GAU) allowed growth rates almost as high as normal.
Note that ThyA mutant suppressor activity (i.e., the ability to complement the ThyA mutations) is a very sensitive assay: it only takes a small number (~100) “correct” ThyA molecules (i.e. molecules with a Cysteine at the mutated site) in a cell for it to grow in the absence of thymidine. So what is limiting the the suppression rate from the mutant tRNAs? What's going on is probably that the mutant Cys-tRNAs, although highly expressed, were poorly loaded with Cysteine (other experiments cited in the paper show that Cys-tRNA variants similar to the low-suppressors here are loaded with Cys 2000-fold less efficiently than the wild-type Cys-tRNA). In other words, while the mutant tRNAs were expressed, not nearly a significant fraction of the relevant codons were being replaced by cysteines, the remainder being used by the correct aminoacids Ile or Met, from the wild-type tRNAs which are still present in the cells.
If this is a reasonable characterization then I was wrong to interpret this paper as supporting my hypothesis (that some codon reassignments are likely to be neutral). I was, however, not able to follow all your reasoning here. Perhaps you can expand on this. Especially the first paragraph. There is a lot packed in there and it is not clear on how it all follows.
[ 11. September 2004, 19:10: Message edited by: Cornelius G. Hunter ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 11. September 2004 16:23
quote:
In trying to keep the jargon down I used the phrase "activate the code change" (as in activate the DNA code change) to describe the process of creating the code change, as distinguished from the question of how well the organism performs once the code change is in place. ....
A little less clear to me is your statement:
quote:
--------------------------------------------------------------------------------
No. The code change is not caused by the Cys-tRNA synthetase gene, but by the mutant Cys-tRNA genes.
--------------------------------------------------------------------------------
Of course that is the case. If you thought I said otherwise then I can see how you would think I'm greatly misinterpreting the paper. But since it is clear that I did not intend to say this, may I suggest a bit more charity in your reading.
I thought that was what you tried to say here: quote:
But the researchers then provided another gene that provides for an alternate DNA code with a single codon change (the gene codes for the cysteine tRNA synthetase, but the translator codon is switched from cysteine to methionine or isoleucine).
Again, it is not the cysteine tRNA synthetase gene that "provides for an alternate DNA code with a single codon change", it is the mutated Cys-tRNA genes that do that. Cys-tRNA synthetase is the enzyme that loads cysteines onto tRNAs.
If I misinterpreted, though, I am sorry. quote:
Regarding my one-on-one basketball game, you say it is wrong because it is really many cells vs. one cell. But of course, it is obvious that I used the analogy to represent the strain types, not the cell counts, and that I was talking about a few vs many situation in terms of cell count. (Nonetheless, I do agree that the idea was a stretch.)
I understood what you meant, but I was pointing out that it was a really bad analogy. If you had a forcibly small bacterial population (say, n=2, as in one-on-one), then indeed it is possible by genetic drift alone that even a lower fitness variant will take over the population (win the game). But in a large population, you would start with a single muant guy and that guy would essentially have to repeatedly, consistently outperform all the others in the culture, in order to take it over in 500 generations. That's a heck of a lot of lucky three-pointers. quote:
Regarding Figure 1 and fitness, you overstate my claim. I said Figure 1 does not indicate that all the codon reassignments are less fit (it does indicate that some are less fit at certain temperatures, but in other cases we cannot draw this conclusion). Of course, this experiment provides limited information, but you are incorrect that it "has nothing to do with fitness." When the results show that a codon reassignment has slower growth rate at certain conditions, then it is a safe bet that it has lower fitness in those conditions. I think what you meant to say is that Figure 1 does not indicate that any codon reassignments, at any temperature, has not lost fitness (in other words, Figure 1 does not reject the possibility that all the codons are less fit). Agreed, but this does not contradict my conclusion.
Well, it makes your conclusion unwarranted - you can't conclude what the fitness of an organism is, based on a short term growth rate experiment. [By the way, that would be also true if an organism had slower growth rate - what if it did grow slower, but also produced a bacteriostatic compound capable of suppressing growth of its competitors?] It's just not the right experiment. quote:
You also said that the researchers examined cell growth at a higher temperature (42C) "because any stress on protein structure caused by the replacement of aminoacid residues is more apparent at that temperature." Actually, this is not necessarily true. In fact, the researchers stated reason was simply because earlier research had shown that "Misincorporation of amino acids through erroneous codon reading is known to cause heat sensitivity in microorganisms."
That's the effect. Heat sensitivity in turn is caused (at least, in large part) by protein instability. It is actually well known that proteins with misincorporated amino acids are more sensitive to heat denaturation. Thus, increasing temperature just measures the cumulative toxicity of amino acid replacements on the proteome. quote:
I concluded that "The Figure 2 results show that the code change is stable," and you replied "Wrong. It shows that it is stable only if it is positively selected." But, again, since I explained that several times, it should be obvious that that was implied. Obviously, my point was not that there wasn't positive selection, but that the system worked at all.
Fine - but then you have to spell out the evolutionary implications of this. A code change is not stable at all, unless it is selected for. quote:
Then you end with some sarcasm about lawyers (violating your own board-related comment from earlier) and getting ID funding (which makes no sense since ID proponents, if anything, would disagree with my hypothesis).
Actually, that wasn't sarcasm (well, just a little, about the lawyers ![[Wink]](wink.gif) ). I have often said on this board and others that it is a shame that the DI wastes its money on political activities and lawyers, when there are in fact people willing to do research in ID-relevant areas. If I were an ID supporter, I'd be upset. What I am serious about is that you have an idea here that can be tested experimentally: that at least some genetic code variants should in fact be more frequent than they seem to be in nature. I don't know whether your hypothesis and the results from testing it are going to ultimately support ID, or what (it would take much more than one set of experimetns to tell, anyway) but sure as heck it's a better use of money than organizing yet another conference in which the same people will talk to each other, or lobbying politicians to change textbooks. Don't you agree? Seems like a no-brainer to me (but then again, I'm a scientist, not a lawyer or a politician). quote:
Having said all that, and while I do agree with you that there are some interesting possible research questions here, I do see your point that the paper does not provide the kind of support I was thinking of. The key point you make, which I misunderstood, was that the cysteine replacements (i.e., the DNA code change) are quite partial. My understanding was that it was pretty efficient, but you make the following point:
quote:
--------------------------------------------------------------------------------
However, even in the presence of Cys-tRNAs, the replacement of Cys at the mutant codons is just partial. This is highlighted by the fact that of the 4 mutant tRNAs expressing strains, only one (Cys-tRNA/GAU) allowed growth rates almost as high as normal.
Note that ThyA mutant suppressor activity (i.e., the ability to complement the ThyA mutations) is a very sensitive assay: it only takes a small number (~100) “correct” ThyA molecules (i.e. molecules with a Cysteine at the mutated site) in a cell for it to grow in the absence of thymidine. So what is limiting the the suppression rate from the mutant tRNAs? What's going on is probably that the mutant Cys-tRNAs, although highly expressed, were poorly loaded with Cysteine (other experiments cited in the paper show that Cys-tRNA variants similar to the low-suppressors here are loaded with Cys 2000-fold less efficiently than the wild-type Cys-tRNA). In other words, while the mutant tRNAs were expressed, not nearly a significant fraction of the relevant codons were being replaced by cysteines, the remainder being used by the correct aminoacids Ile or Met, from the wild-type tRNAs which are still present in the cells.
--------------------------------------------------------------------------------
If this is a reasonable characterization then I was wrong to interpret this paper as supporting my hypothesis (that some codon reassignments are likely to be neutral). I was, however, not able to follow all your reasoning here. Perhaps you can expand on this. Especially the first paragraph. There is a lot packed in there and it is not clear on how it all follows.
I realize now that that paragraph is really badly written, sorry. It should have said something like this:
"However, even in the presence of Cys-tRNAs, the replacement of Cys at the mutant codons is just partial. This is highlighted by the fact that of the 4 mutant tRNA-expressing strains, only one (Cys-tRNA/GAU) grew at rates comparable to normal under thymidine withdrawal conditions."
That is, 3 of 4 of the tRNAs are really weak suppressors of the ThyA mutation, because they can't even generate 100 "corrected" ThyA molecules in the cells (out of probably many thousands synthesized by the 200-copy number plasmid bearing the mutant thyA gene). The same can also be seen for Cys-tRNA/UAU in the missense suppression of mutant AmiE experiment (table 2).
Do you understand why the replacement is always partial, right? Because all these strains still express wild-type tRNAs, which compete with the mutant ones. In addition, 3 of the 4 mutants, including Cys-tRNA/UAU (the one used in the fitness experiment) load really poorly (2000-fold less, probably), and in all cases co-expression of Cysteine-tRNA synthetase (which improves loading and therefore replacement) increases toxicity. It would have been nice to somehow quantitatively measure replacement rates, I guess, but they didn't (I am not even sure how one could do that - maybe by measuring specific activity of missense-suppressed thyA?).
Take it easy.
[ 11. September 2004, 16:38: Message edited by: charlie d. ]
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 11. September 2004 19:07
Charlie:
Ah -- no apology needed, my mistake. Now that I actually read what I typed, you were absolutely right that my description of their process was confusing. I inadvertantly inserted the word 'synthetase.' As it stands the sentence makes no sense because the first half erroneously includes the word 'synthetase' but the second half (correctly) refers to the tRNA anti codon. Thanks for the correction.
Thank you for clarifying your paragraph on partial replacement. That makes sense. And yes, I certainly agree with your sentiment about spending money on science rather than politics.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 11. September 2004 21:30
You're welcome. Hope this discussion helped to clear some of the misunderstandings about genetic code variants, their origin and evolution, and the degree of code diversity one would expect to encounter under the common descent model.
Remember, I'll be watching...
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 11. September 2004 23:21
Charlie:
I appreciate your inputs and hope you do keep watching. In fact, if you care I'd be interested to hear your thoughts on the more general question of whether or not the universality of the DNA code is strong evidence for common descent. Earlier, you made the point that evolution of the code is likely to be difficult because proteome-wide changes are likely to be deleterious at least somewhere in the proteome. Hence, you concluded that code changes should be rare.
Unfortunately, "rare" is hard to quantify. What if there were twice as many variants as we observe? Thrice? After all, it is possible that more variants could be yet discovered. I have a hard time believing such findings would be viewed as problematic for common descent. And we know that zero variants was not viewed as problematic. So we know that anywhere from zero up to some not insignficant number of variants would not be problematic.
What about larger numbers yet? Could unusual population dynamics and environmental shifts be used to explain yet more variants? What if all the plants had a different code? Would that really be a problem? I'm not saying we can take this to the extreme of every species having its own code. But it is not clear to me that common descent is tightly constrained either.
This is highlighted by the fact that the code had to have evolved. Both the complexity and the structure of the code (similar codons tend to code for similar amino acids) suggest that this evolutionary process was involved, involving several stages or intermediates. If the code evolved through such a process, then why is it compelling that the code can no longer evolve much? Of course we can speculate about how the level of complexity finally reached a level where futher change became intolerable, but we can equally well speculate that this level of complexity is well beyond that. Any thoughts?
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 12. September 2004 00:23
quote:
I think that it might be legitimate for Rosenhouse et al to cite the universality of the DNA code as powerful evidence of common ancestry, while keeping OOL (origin of life) separate. This is because DNA code universality could illustrate the common descent of all existing organisms from the first microbe to use the DNA code, without regard to how that microbe came to be.
Darel:
But what if common ancestry can explain a variety of codes? Is the common ancestry model really limited to a universal code with only a handful of variants? If not, then the universality can hardly be powerful evidence of common ancestry. Remember, the common ancestry model can explain a lot of changes, like echolocation and, as you point out, vision.
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 12. September 2004 11:06
quote:
Unfortunately, "rare" is hard to quantify. What if there were twice as many variants as we observe? Thrice? After all, it is possible that more variants could be yet discovered. I have a hard time believing such findings would be viewed as problematic for common descent. And we know that zero variants was not viewed as problematic. So we know that anywhere from zero up to some not insignficant number of variants would not be problematic.
What about larger numbers yet? Could unusual population dynamics and environmental shifts be used to explain yet more variants? What if all the plants had a different code? Would that really be a problem? I'm not saying we can take this to the extreme of every species having its own code. But it is not clear to me that common descent is tightly constrained either.
I think you are asking too much. First, let's make clear that the prediction of rare genetic code variants comes not from evolutionary biology, but from molecular biology.
Let's look first at what we do know, shall we?
1. We know that code variants arise spontaneously at detectable frequency, at least in some classes of organisms (see missense and nonsense suppressor mutations in bacteria), and how they occur;
2. We know that in all cases we have looked at, organisms bearing these spontaneous code variants have lower fitness than wild-type;
3. We know this lower fitness depends on the compound toxicity effects of extensive (proteome-wide) amino acid substitutions on protein structure and function;
4. We know that in order to maintain such code variants with lowered fitness, strong positive selective pressures must therefore be applied (eg, suppressor effects).
Therefore, we could predict that, given a universal genetic code with the molecular genetic characteristics mentioned above, code variants would be rare.
How rare? That's hard to tell, really. We do not know how many variants are even tolerable, as opposed to lethal, or how often missense/nonsense mutations occur that can effectively be suppressed by code changes. Indeed, only relatively a minuscule number of species have been closely analyzed in their genetic code properties (and are they even representative of the whole biosphere?). Therefore, I doubt it would make any real sense to throw out precise-sounding predictions that in fact are just wild guesses.
One could probably say that, whatever the frequency, code shifts should be expected to occur more often in organisms with limited proteomes (in which toxicity effects should be smaller), or organisms living in peculiar environmental conditions, or with unique biochemistries (because we know that temperature and chemical environments may themselves affect protein stability, amino acid availability and usage, etc). But that's pretty much it, really.
quote:
This is highlighted by the fact that the code had to have evolved. Both the complexity and the structure of the code (similar codons tend to code for similar amino acids) suggest that this evolutionary process was involved, involving several stages or intermediates. If the code evolved through such a process, then why is it compelling that the code can no longer evolve much? Of course we can speculate about how the level of complexity finally reached a level where futher change became intolerable, but we can equally well speculate that this level of complexity is well beyond that. Any thoughts?
Yes, of course the code must have evolved. But, as you say, we also know that it must have evolved in organisms that were considerably simpler than those we have now, with much more limited proteomes, different amino acid biochemistries, etc.
Again, we can only speculate, as far as the current code goes, based on what molecular genetic evidence tells us about today's organisms and their codes.
So, going back to common descent, it is true that CD theory would have not necessarily predicted a universal code, but that's because when the theory was formulated people had no idea of the genetic code to start with. Had molecular biology discovered a much more plastic genetic code than what we have, CD would have been perfectly comfortable with extensive genetic code variations (especially if arranged in a nested hierarchy!).
On the other hand, given the type of genetic code we do have, and given its empirically determined constraints on evolvability, the picture that emerges of an extensively conserved universal genetic code with a few sporadic, very limited variants is pretty much what one would expect under common descent.
It could well have been different: we could have had a very flexible code at the molecular genetic level, and found no code variants at all in any organism, or we could have had a highly rigid code like the existing one, but an extensive, wild diversity of codes. Either possibility would have required some additional explanations and theory adjustements, which would be hard to imagine right now, in order to be compatible with common descent. (In fact, at the extreme, the finding of a code with the characteristics of the current one, coupled with extensive, non-hierarchical code variations between organisms, would have pretty much invalidated common descent, I think)
But that was not the case - evidence on the molecular biology of the code ended up fitting rather easily and comfortably under the common descent blankie.
[ 12. September 2004, 11:17: Message edited by: charlie d. ]
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 12. September 2004 19:09
Charlie:
Good points, however, the reason that it seems I am asking for too much is that a strong claim has been made. The universality of the DNA code is claimed to be powerful evidence for common descent, so it is fair to ask for justification. The code cannot be powerful evidence if the existing code (and variants) is one of several situations that common descent can explain.
I think that my questions, rather than asking for too much, instead reveal that too much has been claimed. We know that common descent can explain a fairly significant range of possibilities. How significant? I agree with you that we just do not know precisely:
quote:
I doubt it would make any real sense to throw out precise-sounding predictions that in fact are just wild guesses.
And, as you point out, we also do not know much about the code's evolution. You say that the organisms were likely to be considerably simpler. Actually, that in itself is an interesting question. Woese and co-workers have argued for years that we need a different sort of process, where massive horizontal trading of genes takes place as the three branches of life emerge; otherwise, you need a highly complex common ancestor of the three branches.
But in any case, even with simpler organisms, I think you will have a difficult time justifying the claim that our highly complex code arose in vastly simpler organisms. Perhaps you were not suggesting that, but regardless, the point is one cannot get around the fact that a complex code reflects a complex biochemistry. Going back just one step in the code's evolution, you have something that looks pretty close to today's code, with pretty complex organisms. Yet that final step must have been possible. The code must have been evolvable at that point. Yet after that last step, it was hardly evolvable. The challenge for evolutionists is to explain why this must be so.
The bottom line is that the DNA code we do have fits nicely into evolution, but evolution could also explain a significant range of other situations (given today's empirical knowledge). There are all kinds of explanatory mechanisms that could be used to justify much greater variation, and our empirical knowledge is fairly limited today. This situation does not make for "powerful evidence." You make the point that things could have been different:
quote:
It could well have been different: we could have had a very flexible code at the molecular genetic level, and found no code variants at all in any organism, or we could have had a highly rigid code like the existing one, but an extensive, wild diversity of codes. Either possibility would have required some additional explanations and theory adjustements, which would be hard to imagine right now, in order to be compatible with common descent. (In fact, at the extreme, the finding of a code with the characteristics of the current one, coupled with extensive, non-hierarchical code variations between organisms, would have pretty much invalidated common descent, I think).
Taking your final, parenthetical, point first, I have to wonder what you mean. We currently have that very situation, except that it is not extensive. There are several non hierarchical variants observed and no one is seeing a problem. Are you suggesting that if we tossed in a bunch more variants there suddenly would be a problem for common descent? I don't think so.
Consider an experiment to reverse one of the known DNA code variants, switching it back to the standard code. If it worked, it would be viewed as an example of a flexibility in the code. If it did not work, then it would be explained that after the code change had occurred via evolution, subsequent evolution has now made the reversion unworkable.
Next, as for your hypothetical case of a flexible code without variants, this sort of evidence has not bothered evolutionists in the past. I don't see why it would bother them now. There are plenty of examples, at both molecular and morphological levels, where we find similar designs yet no apparent necessity. While such necessity may be discovered in the future, the point is our current empirical knowledge indicates no such necessity. Histone H4 proteins, for example, are highly conserved yet mutations in their sequences are well tolerated. Or, functionless segments in mouse and human genomes are highly conserved.
Finally, as for your hypothetical case of a highly rigid code with an extensive variety of codes, this too could easily be explained as the result of divergences in the early stages of evolution when the code was still evolving. Of course, as I said earlier, extensive code differences in cousin species would be hard to explain. I'm not saying common descent can explain every situation.
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 12. September 2004 21:09
Open question for anyone,
This may be a shot in the dark, but I'm genuinely curious.
Do we believe the variant codes indicate close ancestry or do we see identical variant codes emerge in unrelated lineages.
In other words, do we see evidence of convergence (even a sub-optimal or error convergence) in the codes?
Salvador
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 12. September 2004 21:26
quote:
Do we believe the variant codes indicate close ancestry or do we see identical variant codes emerge in unrelated lineages. In other words, do we see evidence of convergence (even a sub-optimal or error convergence) in the codes?
Salvador:
There definitely is convergence in the variants. That is, the same variant must have occurred multiple times.
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 13. September 2004 03:19
This post responds to Professor Rosenhouse's Tuesday, September 07, 2004 blog entry at:
http://evolutionblog.blogspot.com/
In this blog entry, Professor Rosenhouse makes some points that would be interesting to discuss. Unfortunately they are obscured by the strawmen arguments that underlie practically all of Rosenhouse's points.
Rosenhouse's strawmen all follow the same pattern. It is a common pattern in these discussions, and goes like this:
1) Skeptic is told X is powerful evidence for the theory.
2) Skeptic explains why X is not powerful evidence.
3) Skeptic is told that the theory can explain X so therefore skeptic has failed to falsify the theory. Point #2 is obviously invalid and this reveals that skeptic fails to understand the theory.
In this strawman, the skeptic's point is overstated. His explanation that X is not powerful evidence is interpreted as an attempt to falsify the theory. Since the skeptic failed to falsify the theory, he must be all wrong, and the ridicule follows. Here's how Rosenhouse uses the strawman three times in the blog.
First, drawing on Ken Miller, he claims that the variants in DNA code are powerful evidence for common descent. After I explain that this is not the case, Rosenhouse invokes the strawman argument, placing me in the position of attempting to "condemn" common descent, and failing to fully appreciate the theory:
quote:
This is another example of Hunter being too enamored of the idea that convergence is some sort of problem for evolution. He seems to think that the mere fact that same codon shift was converged upon in separate lineages condemns common descent. He consistently fails to consider what sorts of natural mechanisms can account for such convergences.
Of course, I never said anything about the evidence condemning the theory. Next, Rosenhouse has claimed that the fossil record is strong evidence for evolution. True, the fossil evidence provides evidence, but I also point out that the evidence has substantial problems and must be seriously caveated. The fossil record has substantial gaps and convergences. And when something as phenomenally complex as the trilobite eye appears abruptly in the fossil record, that is not exactly "powerful evidence" for evolution. Evolution has no explanation of how it could have arisen.
So again, Professor Rosenhouse places me in the position of arguing that it is impossible for such complexities to arise via evolution. He argues that "A process in which random variations are sifted through a non-random selection process can lead to outcomes far more complex than what you started with," as though I had said it cannot. The problem is not that complexities falsify evolution, the problem is that complexities caveat the evidence for evolution. We cannot simply ignore the fact that evolution cannot explain how these designs arose. Rosenhouse wants to obviate the problem by pointing out that degrees of complexity are difficult to define. Agreed, but that misses the point. The complexities in question here are beyond the explanatory power of evolution, that is the point. Perhaps someday in the future this will change, but that too misses the point. We are evaluating the theory today, using today's knowledge base, not some projection of what future experiments will reveal.
Professor Rosenhouse continues with the strawman, saying that "To determine if a particular system could [or could not] have been crafted by natural selection it is not enough to diagnose it as being complex." True enough, but so what? I did not say that all complexities defy natural selection.
Next, Rosenhouse claims that comparative anatomy is strong evidence for evolution. I explained that if similar designs are present in distant species, where common descent cannot be used to explain those similarities, then common descent need not be invoked to explain similarities in species that are not so distant. Homologies, such as the pentadactyl pattern, were a key argument for Darwin. He viewed them as a mandate for common descent. But this argument is contradicted by the convergences.
And once again, Rosenhouse uses a strawman, placing me in the position of arguing that convergences argue against evolution, rather than against the homology evidence (note: convergences contradict the evolutionary idea of contingency so in that sense they argue against evolution; but not because evolution cannot explain them). He says that "there is nothing in the mere fact of convergence that calls evolutionary theory into doubt," as if that is what I had argued.
These strawmen arguments make the debate tedious, and they overshadow relevant issues. For instance, regarding comparative anatomy Rosenhouse provided an argument independent of the strawman. He claimed that homologies are good evidence because they are sufficiently different from analogies. Specifically, Professor Rosenhouse claimed that with homologies "The level of morphological similarity here is far greater than in any of the examples of convergence." Will Rosenhouse agree the evidence is weak if this is not so? Well this certainly is not so. Just look at the pentadactyl pattern examples -- the geometries, aspect ratios, and so forth show radical differences. One can hardly see the correspondence in some cases. Meanwhile, there are many convergences that are quite similar. And if this isn't enough for Rosenhouse, he still has not explained how similarities in cousin species can arise from different development pathways.
[ 14. September 2004, 01:42: Message edited by: Cornelius G. Hunter ]
IP: Logged
|
|
Cornelius G. Hunter
Member
Member # 81
|
posted 14. September 2004 04:23
This post responds to Professor Rosenhouse's second blog entry of Tuesday, September 13, 2004 at:
http://evolutionblog.blogspot.com/
I explained that complexities that evolution cannot explain are a problem for evolution. Such complexities are observed in the fossil record as well. I noted the trilobite eye as an example. Professor Rosenhouse disagrees. He says that "the fossil record provides strong evidence for common descent. No caveats. No ambiguity." But how can this be if evolution cannot explain how structures such as the trilobite eye arose? For Professor Rosenhouse, there is no problem. My assertion that evolution has no explanation for the trilobite eye is "simply false." Why? Because, as he writes:
quote:
As I pointed out earlier, the period between the origin of multicellular life and the appearance of the eye spans 200 million years. That's plenty of time for natural selection to produce all sorts of complexity.
So there we have it. This is what passes as a scientific explanation in evolution. It just happened -- somehow. This finding appears out of nowhere in the fossil record. The trilobite eyes were perhaps the most complex ever produced by nature [1], and one expert called the design "an all-time feat of function optimization" [2]. And though evolution cannot explain how this design arose, Professor Rosenhouse nonetheless tells us that evolution can explain how this design arose. The explanation is that there were all those millions of years in which they could have developed. In the hands of evolutionists, science becomes nothing more than unfounded speculation.
Notes
1. Lisa J. Shawver, “Trilobite Eyes: An Impressive Feat of Early Evolution,” Science News, p. 72, Vol. 105, February 2, 1974.
2. Riccardo Levi-Setti, Trilobites, Second Edition, pp. 29-74, The University of Chicago Press, 1993.
[ 14. September 2004, 23:36: Message edited by: Cornelius G. Hunter ]
IP: Logged
|
|
Darel R. Finley
Member
Member # 100
|
posted 15. September 2004 11:20
quote:
But what if common ancestry can explain a variety of codes? Is the common ancestry model really limited to a universal code with only a handful of variants? If not, then the universality can hardly be powerful evidence of common ancestry. Remember, the common ancestry model can explain a lot of changes, like echolocation and, as you point out, vision.
Dr. Hunter,
I agree that homology and analogy cannot count as evidence of Darwinism unless it can be shown that Darwinism makes some specific predictions of what kinds of homologies and analogies will and will not be observed -- and that those predictions are not contradicted by known evidence.
The reason I brought up Behe was to point out that the same problems with OOL that cause many evolutionists to consider it a separate subject are present in explaining many steps up the microbe-to-human ladder as well, so evolutionists who cede OOL as a mystery while saying that microbe-to-human is not, are making a mistake.
However, in your original message at the top of this thread, you say
quote:
Rosenhouse tries to avoid all this by saying that, properly understood, the evolution of the code is not a part of evolution, but rather the origin of life (OOL) problem. So why does Rosenhouse draw the evolution / OOL line where he does? Does he see a genuine distinction, such that in the former designs arose via evolution, but in OOL the designs are likely not to have evolved? If this is the case then why not simply say so?
I think that it is fair for evolutionists to draw a distinction between OOL and the microbe-to-human transformation, because their alleged evolutionary mechanism (generation-to-generation variations filtered by natural selection) requires there to be reproducing organisms, and before OOL there were no reproducing organisms.
IP: Logged
|
|
Salvador T. Cordova
Member
Member # 959
|
posted 15. September 2004 22:15
Dr. Hunter,
There is a old paper from the now famous (infamous) Pajaro Dunes conference. I provide it here with the hope it might further our inquiries.
Paul Nelson on Universal Codes
Some particularly pointed quotes:
quote:
Researchers now expect to encounter further variants. “It seems obvious," Caron argues, "that the number of cases of deviations observed will increase rapidly in the future ."[43]
and
quote:
Yet variants in the nuclear code discovered more recently, are, Fox argues, of a different order: "Some 'real' (nuclear] exceptions have come to light in both eukaryotic and prokaryotic free-living organisms, and the notion of universality will have to be discarded."
For instance, "in at least four species of ciliated protozoa, the codons UAA and UAG [stop codons in the universal code] occur in nuclear genes and are translated as Gln during cytoplasmic protein synthesis."[41] In the bacterium Mycoplasma capricolum, UGA encodes Trp, rather than termination (stop) as in the universal code.[42] Other variants are given in the Jukes and Osawa article.
Oh, I found this juicy quote:
quote:
Carl Woese
We cannot expect to explain cellular evolution if we stay locked in the classical Darwinian mode of thinking....The time has come for biology to go beyond the Doctrine of Common Descent.
![[Cool]](cool.gif)
PS
for the sake of interested readers, the issues with the genetic code have dominated this thread, however, that is only the tip of the iceberg. The issue of different devlopmental pathways arriving at the same structure is a large topic.
For the readers, I found think this article by ISCID fellows Jonathan Wells and Paul Nelson : Homology a Concept in Crisis is a good summary.
[ 15. September 2004, 23:38: Message edited by: Salvador T. Cordova ]
IP: Logged
|
|
|
Printer-friendly view of this topic