|
Author
|
Topic: "Junk DNA" as Cannon Fodder
|
John Bracht
Member
Member # 5
|
posted 03. September 2002 19:51
This is just a bit of musing that occurred to me as I was thinking about mutation rates the other day, and suggests to me a possible function of junk DNA.
My logical train of thought was as follows: the number I usually see cited for the background mutation rate in a genome is usually around 1x10^-9 per base pair per generation. In other words, the chance that a given base will mutate in any given cell division (generation) is around one in a billion. Since there are around a thousand base pairs per gene, the probability that any given gene will get a mutation (one of its base pairs will mutate) is around 1x10^-6 per generation.
The interesting thing is that there are only around 30-60,000 genes (estimated) in the human genome, while there are 3.1 billion total base pairs (much of the genome is non-coding "junk"). That means that the probability of getting a mutation in any gene in the genome (per generation or cell duplication) is around 0.045 ((1x10^-6)*(45,000; an average of 30,000 and 60,000)). This means that we would expect, on average, to see 4.5 mutations in 100 organisms (in genes), per generation. However, doing a similar calculation for the whole genome (genes plus junk DNA) we should see 3.1 mutations per organism per generation overall.
What does this mean? The number of mutations in genes is about 100 times lower than the number of mutations overall in an organism. It seems that the non-coding DNA is, in some sense, "drawing the fire" in such a way that the probability of mutating the "junk" is quite high--but the probability of hitting an essential gene-coding section is quite low.
Of course, there is the subtlety that many mutations arise when DNA is replicated and the total number of these mutations is dependent only on the total amount of DNA being replicated. So if the "junk DNA" wasn't there, the total number of mutations (of this type) would be drastically reduced, to around the same level we see in genes alone.
However, there are some sorts of mutations that do not come from replication but rather come from chemical or radiation mutagenesis, and it is these sorts of mutation events that penetrate the nucleus and "hit" the DNA. These types of mutations are not dependent on total DNA, and it occurs to me that, given the total amount of non-coding DNA in a cell, the probability of such a mutation event hitting a gene is very, very low. Indeed, the probabilities are good (around 100 times greater) that the mutagen will damage only the "junk" DNA instead of coding segments.
The best analogy I can come up with is a game I used to play with friends in college. It's called "Axis and Allies" and it's a Risk-like strategy game based on WWII. In the game, there are different values to different pieces, and corresponding different offensive/defensive strengths. In a battle, one wants to bring "cannon fodder" or lower-valued pieces, along with one's stronger pieces, so that those lower-valued pieces can take the damage while the stronger ones are preserved.
The analogy with "junk" DNA and coding DNA is obvious. The "Junk" DNA may function as cannon fodder to take the hits from mutagenesis, thereby protecting the precious coding segments. In this case, the function of the noncoding DNA is quite subtle--it doesn't produce proteins or affect gene regulation, etc; indeed, its function is utterly disconnected from any actual sequence and relies instead entirely upon its sheer bulk. And cells have plenty of it.
This makes good design sense. If a designer is planning for life to exist in a world where mutations occur (a world like ours), it makes sense to incorporate safeguards against mutagenesis damage. Utilizing "junk" DNA, in a correct ratio relative to coding DNA, would be an excellent way to do this. An interesting design prediction would suggest that the amount of "junk" DNA in cells should be at or near an optimal value for protecting the genome from degradative mutations. Testing this hypothesis would be as simple as removing "junk" DNA from different organisms and testing their survivability in environments with enhanced mutagenesis levels. Or, one could insert extra "junk" DNA into cells and test for enhanced resistance to mutation (of course, there might be other trade-offs to consider, like the cost of carrying around and replicating too much non-coding DNA). Any comments?
John Bracht
IP: Logged
|
|
Art
Member
Member # 179
|
posted 03. September 2002 23:01
quote:
This makes good design sense. If a designer is planning for life to exist in a world where mutations occur (a world like ours), it makes sense to incorporate safeguards against mutagenesis damage. Utilizing "junk" DNA, in a correct ratio relative to coding DNA, would be an excellent way to do this. An interesting design prediction would suggest that the amount of "junk" DNA in cells should be at or near an optimal value for protecting the genome from degradative mutations. Testing this hypothesis would be as simple as removing "junk" DNA from different organisms and testing their survivability in environments with enhanced mutagenesis levels. Or, one could insert extra "junk" DNA into cells and test for enhanced resistance to mutation (of course, there might be other trade-offs to consider, like the cost of carrying around and replicating too much non-coding DNA). Any comments?
The "experiment" has beend one (many times, in fact). Bacteria are fairly devoid of "junk DNA", and survive mutagenic insults that most multicellular organisms cannot tolerate. Saccharomyces cerevisiae is also fairly free of junk DNA, yet persists quite nicely, even when grown in rather noxious environments. Arabidopsis possesses a tiny fraction of the junk DNA that is seen in most plants (and animals, for that matter), yet it is at least as hardy as even closely related, "junky" plants.
To name three examples, off the top of my head.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 03. September 2002 23:41
As Art has pointed out it seems that 'junk DNA' is not really needed but I would still like to comment on John's proposal.
John suggests that this would be a design prediction. I assume that John both includes natural and intelligent design into the mix. Since the ID inference takes place by showing that the probability for a hypothesis in light of all chance hypotheses is too small, we need to look at whether or not the hypothesis of a Darwinian pathway can be excluded.
But I would also like to approach this from a different perspective, that is, when looking at the genome we also detect pseudogenes. Would such genes be useful for a design hypothesis?
If we were interested in finding positive evidence of ID, we need to better understand motive or goal(s) of the intelligent designers before we can determine if junk DNA has been designed. After all why would a designer use junk DNA versus a more active repair mechanism?
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 04. September 2002 12:48
Good morning John,
Here is some research on "junk DNA" that you might find interesting.
Btw, I have never liked that term. I believe it only shows how little we know about the function and design of DNA.
http://www.sciencedaily.com/releases/2002/08/020830072103.htm
Essential Cell Division "Zipper" Anchors To So-Called Junk DNA
PHILADELPHIA - When cells divide in two, they must carefully manage the process by which their DNA is replicated and then apportioned to the daughter cells. In one critical step along the way, the replicated DNA strands - or sisters - are held together for a period by a temporary scaffold of bridging proteins. When the timing is right, the proteins unzip, allowing the DNA sisters to separate. Errors in this or other steps in cell division can lead to cell death, faulty development, or cancer, which is largely defined as misregulated cell division.
Scientists have had a number of questions about these important bridging proteins, called cohesins. For example, how and where do the proteins attach themselves to the DNA? To protect genes from inappropriate activation, DNA is tightly wrapped around small proteins called histones and then further coiled into a higher structure called chromatin that serves as an effective accessibility barrier to the genes.
In a new study in the August 29 issue of Nature, researchers at The Wistar Institute identify a cohesin-containing protein complex that reshapes chromatin to allow cohesins to bind to DNA. In doing so, they also identified the locations on the human genome where the cohesins bind. Somewhat to their surprise, the binding sites were found to be a repetitive DNA sequence found throughout the human genome for which no previous role had ever been identified. These bits of DNA, known as Alu sequences, are liberally represented along those vast stretches of the human genome not known to directly control genetic activity, sometimes referred to as junk DNA.
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 04. September 2002 22:06
Art,
I think you're too hasty in simply dismissing my ideas. You neglect the fact that the bacteria and yeast may have rather different requirements in terms of resistance to mutation (eg, more advanced multicellular life may have lower intrinsic tolerance to perturbations) and thus the different amounts of "junk" DNA in these cells may be a reflection of those differing design constraints. Also, different organisms live in different environments, with differing levels of radiation and other DNA-damaging stresses. Possibly some of the differences in "junk" DNA may reflect those differences.
That's why I recommend removing these confounding variables by creating genetically engineered organisms with artifically altered levels of "junk" DNA to test in high-vs-low mutation environments. This should provide some compelling evidence one way or another (it should be the crucial experiment to test the hypothesis). Perhaps my ideas are wrong--but let's not dismiss them entirely until the requisite test has been performed.
Frances:
You are trying to drag this thread into a lot of absolutely irrelevant areas (intentions of the designer, the question of pseudogenes, etc) and I refuse to be drawn off topic (and I'd appreciate it if you would stay on topic). However, there was one comment you made that was relevant:
quote:
After all why would a designer use junk DNA versus a more active repair mechanism?
(by the way, the only intention we need assign to the designer is that he/she/it wanted to create an organism well-adapted to its environment)
The answer to the question is that organisms do utilize active repair mechanisms. As I understand it, there are at least two specific mechanisms that check the DNA when it's being replicated (one in the replicase, and another entire set of enzymes that checks the DNA after replication). And there are lots of checkpoints during replication that make sure the DNA is undamaged before replication proceeds. These make good design sense. But it also makes good design sense to utilize "dummy" DNA in addition to active repair mechanisms, to absorb the majority of the potentially damaging mutations.
Nobody:
Thanks for the link. I found that report to be rather stunning, to say the least. I've done some research into Alu sequences, and had already learned of about 3 or 4 definite functions these sequence perform, so I'm a little surprised that they were considered "junk" DNA in the article. Nevertheless, it's a remarkable discovery and just hints at the complexity and intricacy of cellular function waiting to be discovered. That certainly sounds like a design-based prediction, doesn't it?
John
[ 04 September 2002, 22:08: Message edited by: John Bracht ]
IP: Logged
|
|
rafe gutman
Member
Member # 134
|
posted 04. September 2002 22:46
quote:
by john bracht:
If a designer is planning for life to exist in a world where mutations occur (a world like ours), it makes sense to incorporate safeguards against mutagenesis damage. Utilizing "junk" DNA, in a correct ratio relative to coding DNA, would be an excellent way to do this.
i have a better idea, how about keeping a spare copy of each gene around? that way, if one copy gets damaged, the cell can use the other copy as a template for repair. this would take up about 10 times less DNA than "junk" DNA would.
IP: Logged
|
|
Art
Member
Member # 179
|
posted 04. September 2002 23:13
quote:
Art,
I think you're too hasty in simply dismissing my ideas. You neglect the fact that the bacteria and yeast may have rather different requirements in terms of resistance to mutation (eg, more advanced multicellular life may have lower intrinsic tolerance to perturbations) and thus the different amounts of "junk" DNA in these cells may be a reflection of those differing design constraints.
But things are really reversed - bacteria, yeast, and other haploid organisms are going to be inherently more sensitive to mutational change (in this simple way of thinking of things) than diploid (or polyploid) organisms. Your ideas would lead one to predict the opposite of what we see in nature, John.
Also, please remember that I had one example from each different "flavour" of living thing - haploid unicellular prokaryote, haploid unicellular eukaryote, and "advanced multicellular" organism.
quote:
Also, different organisms live in different environments, with differing levels of radiation and other DNA-damaging stresses. Possibly some of the differences in "junk" DNA may reflect those differences.
An interesting thought, but one that again is not consistent with what we see in nature. (Living things do have a range of adaptations suited for tolerating different environmental insults, but there is really no correlation between genome size or junk DNA content and such insults.)
quote:
That's why I recommend removing these confounding variables by creating genetically engineered organisms with artifically altered levels of "junk" DNA to test in high-vs-low mutation environments. This should provide some compelling evidence one way or another (it should be the crucial experiment to test the hypothesis). Perhaps my ideas are wrong--but let's not dismiss them entirely until the requisite test has been performed.
IMO, the test has been done. That's why I am not very enamoured with your idea.
Anyways, what experimental system would you imagine to be amenable to the experiment you propose, John? Technically, how do you imagine the experiment could be done?
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 05. September 2002 00:35
John seems to think that I am adding a lot of irrelevant issues to the discussion but I do not believe that they are that irrelevant and let me explain.
John is arguing that 'junk DNA' may be indicative of design. I assume that John includes both natural and intelligent design. Natural design has a promising hypothesis based on survival. So now we need to understand what mechanism(s) intelligent design proposes in this 'shoot off'.
John then states:
quote:
(by the way, the only intention we need assign to the designer is that he/she/it wanted to create an organism well-adapted to its environment)
At least John's accusation that my questions about motives and intention/mechanisms was irrelevant seems to have been contradicted by the claim that he/she/it/they want to create an organism well adapted to its environment. With John's motive we can at least not exclude natural design, but can we include intelligent design? What hypotheses exist that would allow us to differentiate the intelligent designer from the natural designer here?
But things get a bit more complicated, how do we establish what is junk DNA (non-coding) which was added by the intelligent designer versus the natural designer? John seems to suggest that the 'designer' did not add much 'junk DNA' to the bacterium but somehow did insert additional junk DNA into mammals? or just humans? Does John envision an active intelligent designer over the past 4 billion years?
Another issue that I would like to raise is the the initial argument. If the probability of the genome is 10-9 per base pair per generation then how is adding junk DNA going to change this number? Is junk DNA somehow going to attract mutations away from the rest of the genome or, more likely, is the junk DNA going to mutate at the same rate as the rest of the genome?
There seems to be something wrong in the initial calculations imho. How can a constant per base pair mutation rate suddenly become less for non-junk DNA? If I have a chance of 1 in a million to be hit by a car, how does having a million people around going to change MY odds? In fact there seems to be little reason to assume that a longer DNA is going to be less affected by a mutative agent.
Perhaps John can explain this detail further since it seems to get to the crux of his argument of 'cannon fodder'.
There are some additional issues which seem to point to a designer in junk DNA, namely the 1/f distribution which is typical of evolving systems for instance and the Zipf law like distribution
Seems that a natural design inference seems to be the best explanation of the observations of 'optimize survival'. Lacking any additional distinguishing feature for the intelligent designer, it's mechanisms etc, we might not want to jump to a design inference.
Btw to hear DNA/Proteins as composers check out thsi link
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 05. September 2002 00:38
Hi Rafe,
In a sense, the complementary strand of DNA for a gene is a "spare copy" that the DNA repair mechanisms actually use to correct mutations. So your idea is already used, in some sense. And it's worth pointing out that diploid organisms certainly have two copies of each gene, so even if one is knocked out by mutation, the other is often able to compensate (though not always). So your ideas are like Frances's; they have already been implemented, but I am proposing another DNA protection mechanism on top of these pre-existing mechanisms.
Art:
quote:
But things are really reversed - bacteria, yeast, and other haploid organisms are going to be inherently more sensitive to mutational change (in this simple way of thinking of things) than diploid (or polyploid) organisms. Your ideas would lead one to predict the opposite of what we see in nature, John.
Really? I said that the reduced complexity of bacteria and yeast suggests they may be more robust to perturbation (mutation) and may not need the advanced protection required for more complex life-forms (multicellular organisms, etc). Haploid vs diploid didn't enter into my calculations. I have to admit, my ideas about lower complexity being more robust to mutation are more intuitive than anything empirical, but it seems right. After all, if you have fewer total genes (like bacteria and yeast) then you won't have as many mutations, period. That means you might not need as much "gene protection". Furthermore, it just makes sense that a simple machine (like a screwdriver) will be more robust to dings, bangs, being dropped, etc., than will a more complex one (like a computer).
quote:
(Living things do have a range of adaptations suited for tolerating different environmental insults, but there is really no correlation between genome size or junk DNA content and such insults.)
Not that I don't trust you, but do you mind backing up this blanket statement with some evidence? You're just insisting that there is no correlation between genome size and environmental insults; but without some evidence I see no reason why I have to take your word for it. In fact, I've read that there IS a correlation between cell size (physical size) and "junk" DNA content. Here's the reference:
Beaton M, Cavalier-Smith T. Eukaryotic Non-Coding DNA is Functional: Evidence from the Differential Scaling of Cryptomonad Genomes. Proceedings of the Royal Society of London B 1999;266:2053-9
It has been hypothesized that the "junk" is functioning as a stuffing (making more room for transcriptionally required enzymes to access the genes), but my hypothesis is that these physically larger cells are usually found in more complex organisms and that they require more mutation protection, which is provided by the "junk" DNA. So the evidence actually supports my ideas, as far as I can tell, and not your assertions. But provide the citations, and let's see where the evidence points!
John
P.S. Frances, please go back and read my initial post where I deal with the question of the per-base pair rate of mutation. It's a good point, and I did think about and deal with that objection. If you feel my treatment was unsatisfactory, we can discuss it.
[ 05 September 2002, 00:42: Message edited by: John Bracht ]
IP: Logged
|
|
charlie d.
Member
Member # 159
|
posted 05. September 2002 08:23
John:
I think the most significant evidence against a significant, specific functional role of "junk" DNA is the fact that organisms of very similar "complexity" and biology vary significantly in their genome sizes: if I remember correctly, salamander genomes can vary >100-fold, fish by 10 folds, even mammalian genomes can vary by about 50%. [Edit: values corrected, ain't Pubmed grand]
I also seem to remember than a hypothesis similar to yours had been proposed at some point (though I can't pinpoint a reference), with particular regard to natural chemical carcinogens. I have not heard anything about it in a long time, though, so it probably did not pan out when tested (this may be what Art is referring to).
Contrary to what some evolution critics think and say, there is a lot of interest in non-genic DNA, and molecular biologists have tried and are trying hard to understand whether there is some function for at least some of it, and what it might be. Some evidence exists in this regard, like the paper nobody posted (however, I'd caution that what that paper shows is that the cohesin complex can bind to some, not all, Alus, not that the function of Alus is to bind the cohesin complex).
That such function has not been found yet may mean that it is not that obvious, or that there is none. Time will tell, but at this point I'd say the odds are not that good.
PS: I don't think your proposed experiment can work: we do not have the technology to stuff many times a genome size of DNA into some poor organism, nor to delete most of the "junk" out of others.
[ 05 September 2002, 10:09: Message edited by: charlie d. ]
IP: Logged
|
|
Elend
Member
Member # 326
|
posted 05. September 2002 08:35
John:
quote:
This means that we would expect, on average, to see 4.5 mutations in 100 organisms (in genes), per generation. However, doing a similar calculation for the whole genome (genes plus junk DNA) we should see 3.1 mutations per organism per generation overall.
What does this mean? The number of mutations in genes is about 100 times lower than the number of mutations overall in an organism.
In other words is 3.1 mutations per organisms overall vs. 0.045 per organism per genes. Note that no matter how much you would increase the amount of "junk" you will still have THE SAME number of mutations per organism per genes. Consequently, no junk at all would STILL make it 0.045 per organism per genes. That is if the mutation rate is indeed constant the same as you started the calculus with. So the absolute number of mutations in genes is independent of the junk. Using "junk" DNA as pro-design argument does not stand. If the "junk" DNA would perform as some sort of redundancy code helping to correct the errors in the true genes - that would be a better argument.
quote:
It seems that the non-coding DNA is, in some sense, "drawing the fire" in such a way that the probability of mutating the "junk" is quite high--but the probability of hitting an essential gene-coding section is quite low.
No. The absolute value of mutations is indeed larger in junk DNA simply because is more junk than essential coding genes. The probability of mutation is the same overall - and the same you originally started your computations with. I gotta agree with Frances on this one.
IP: Logged
|
|
Frances
Member
Member # 169
|
posted 05. September 2002 13:25
John, if the overall mutation rate per basepair remains fixed then how could adding more basepairs make a difference? Unless the argument is that mutations have a limited capacity to mutate the genome. Radiation for instance would not qualify as such but perhaps other mutations would? But in general when the mutation rate per basepair is fixed, adding more basepairs is not going to make ANY difference.
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 05. September 2002 13:41
quote:
Nobody:
1. Thanks for the link.
2. I found that report to be rather stunning, to say the least.
3. I've done some research into Alu sequences, and had already learned of about 3 or 4 definite functions these sequence perform, so I'm a little surprised that they were considered "junk" DNA in the article. Nevertheless, it's a remarkable discovery and just hints at the complexity and intricacy of cellular function waiting to be discovered.
4. That certainly sounds like a design-based prediction, doesn't it?
John
1. You're welcome.
2. Good! I'm glad you liked it. It seems that many other people are having a similar reaction. I expect that we will get more stunning reports as DNA gets studied more extensively.
3. I think the term "junk DNA" should be discarded immediately. However, it is my observation that old habits die slowly.
4. Yes. I believe it is now overwhelmingly obvious that we are looking at highly intelligent programming, programming that is far beyond current human capability.
IP: Logged
|
|
John Bracht
Member
Member # 5
|
posted 05. September 2002 19:52
Frances, Elend,
I've already addressed the very issue you are raising. Please do me the very basic duty of at least reading my initial post carefully and thoroughly before responding to my arguments. I won't do your homework for you by explaining my argument again, but I will give you a hint: look carefully at the paragraphs following "Of course, there is the subtlety..."
If you two think that I have not adequately addressed this issue, we can discuss it. But until you actually read and understand my argument, we really have nothing to talk about.
John
IP: Logged
|
|
Elend
Member
Member # 326
|
posted 06. September 2002 03:41
John, I can't speak for Frances, but I can say that I did read your whole post. The fact that I do not agree with your conclusion and handling of math does not mean I did not understand it. Also the fact that some of us even bothered to post comments is a sign that a discussion is needed. Now to concrete issues.
As I understood it, the mutation rate you provided in the begining, 10^-9 per base pair per generation is the total rate, including all possible mechanisms. Did I get this right? Please explain then how adding more "junk" DNA would affect the total number of mutations in the genes. My math says it simply doesn't change.
You also state: quote:
However, there are some sorts of mutations that do not come from replication but rather come from chemical or radiation mutagenesis, and it is these sorts of mutation events that penetrate the nucleus and "hit" the DNA. These types of mutations are not dependent on total DNA, and it occurs to me that, given the total amount of non-coding DNA in a cell, the probability of such a mutation event hitting a gene is very, very low. Indeed, the probabilities are good (around 100 times greater) that the mutagen will damage only the "junk" DNA instead of coding segments.
First, how exactly those types of mutations are not dependent on the total DNA? Longer DNA means larger surface, which means more likely to be hit by a mutation capable radiation. Do you have any reference to a study that supports your assumption? Or you actually say that the mutation RATE is independent of the total DNA? If yes, we are back where we started - how is adding more "junk" DNA affecting the (absolute) number of mutations in the coding DNA?
If you direct me again to read your first post, I will take it as a refusal to continue a discussion and give up posting on this board.
[ 06 September 2002, 03:50: Message edited by: Elend ]
IP: Logged
|
|
|
Printer-friendly view of this topic