How well do we understand DNA?

I'm just an amateur, kind of an "armchair biologist", but my wife is a guraduate student working on developing a genetic basis for the classification into species for different varieties of a common agricultural pest. So hopefully you won't consider my input enitrely without merit. We do have an ability to determine, from a given sequence of nucleotides, what amino acid residues will be connected as a result. That's a pretty simple code. But protein function is a matter of shape, which results from the complicated folding interactions of those residues. While we're working hard on it*, we don't as yet have the ability to predict shape from sequence except for the simplest sequences.The major surprise of the last few years was how simple our DNA really is, less than 25k genes. We had predicted at least 100,000, based on the number of proteins in our body. So clearly the interaction of that DNA with our body is a hell of a lot more complicated than we thought. So, to answer your question, I would say that we understand the function and structure of DNA extremely well, because those things are very simple. What is very, very complicated, and what we don't yet fully understand, is the myriad interactions between genes and the body, and between genes and other genes. DNA accumulates copying errors because copying is never perfect; these errors are largely reduced by error-checking mechanisms in the replication process. But it isn't perfect, and so we know that mutation must occur.Some organisms apparently can react to certain stresses by reducing the effectiveness of these error-checks; as a result the DNA mutates more often. We know that random mutation will always occur because of the second law of thermodynamics. There's no way to completely prevent it. If what you're asking is "can a cell allow certain areas of the DNA to mutate, and completely prevent other areas from doing so?" the answer appears to be "no". Some areas of DNA appear to mutate more frequently than others, but no area of DNA can be perfectly protected from mutation. Biologists hate the term "junk DNA" because it causes precisely this confusion. The sequences you refer to are not transcribed, you are correct.But that's not to say that they're without purpose. As I said earlier, it's impossible to prevent mutation, and a mutation is more likely to be harmful than to be helpful. Having the vast majority of your DNA be sequences that will never be transcribed may help "shield" valid genetic sequences from mutation by making it less likely that a random mutation will occur in the middle of a gene.Moreover, junk DNA has an additional function. If you were a teacher, and you suspected two students had copied each others papers (and not just written two similar papers because they both researched the same topic), you would know if they had done so because both papers had the same spelling mistakes, grammatical errors, and typos. So too do these "plagerized errors" allow us to reconstruct evolutionary histories, because the only reason two species would have the same collection of errors and typos in their DNA is if they had copied their DNA from the same source - that is, they shared a common ancestor.*If you'd like to help out, you can download and run the distributed "Folding @ Home" client, which uses your spare CPU cycles to run protein folding analysis. No webpage found at provided URL: http://folding.stanford.edu/This message has been edited by crashfrog, 01-16-2005 13:10 AM

Technically, no, it's not. That wasn't meant to be a statement of fact, but rather, a segue into how junk DNA is employed in the lab, and how it holds up the evolutionary model. But I'm sorry that it was clumsy and easily misunderstood.

It's simple probability, to me. Lets say that an energetic photon penetrates the cell's nucleus and snips a chromosome in two pieces. The nucleus's repair machinery comes into play but there's always the risk that the DNA won't be repaired quite right at the point where it was broken.If, say, 70% of that chromosome is non-functional, then there's a 70% chance that imperfect repair occurs at a section where it doesn't matter. I hope Rrhain can check my math on this. Lets try and work it out together; I hope you'll interject where you disagree.Lets say that an organism has a genome 100 bp long, and a mutation rate of 1 per hundred base pairs per generation. This organism has no junk DNA whatsoever - every sequence in its genome is either protein-encoding or regulatory. We would therefore expect every generation of this organism to impact a functional sequence.Now lets double the length of it's genome with junk DNA. Every generation should now have two mutations, but the odds that each one will impact a functional sequence are 50%. The outcome that either mutation impacts a functional sequence is three out of four possible outcomes:1) First one "hits", second one "misses"
2) First one misses, second one hits
3) First one hits, second one hits
4) First one misses, second one missesThe way I figure it, that's a 75% chance per generation that a mutation has "hit" a functional sequence, down from 100% for no junk DNA. Unless these outcomes are not equally likely, as I have assumed. If my math is flawed, and it feels like it is, I hope it will be pointed out to me.

Well, there ya go.

Well, I mean, it wasn't even my idea - it was just something I heard from somewhere, and it sounded plausible. Now I know that it isn't. It's not like my scientific career was on the line, because I don't have one.

There's simply no physical way in a thermodynamic universe to prevent mutation in any part of DNA. It can be reduced, and is, but it cannot be prevented completely as you suggest. There's no test that operates on DNA except the expression of genetics in the morphology of individuals who are then selected for or against by the environment. If there were, we could detect the cellular machinery that did so; if you believe it exists its inbumbent on you to prove it, not us to prove you wrong.

Er, no, you have, remember? When you said this?

quote:

---

I can conceive writing a computer program that copies itself and, each time it does, changes certain sections of the program on a random basis. Certain core areas of the program could be identified so that they could be prevented from being changed (or else the whole thing would quit or go haywire).

---

The reason we know that random mutation operates, and that all mutations are not simply pre-programmed randomness restricted to certain sequences, is that all sequences mutate. It's impossible to completely protect a given sequence from mutation. There's no way to distinguish them, because those are exactly the same thing. We find that some genetic locii mutate more frequently than others, due to their chemical structure (some nucleotide sequences are physically less resistant to mutation than others) or their position in the chromosome. This may even be a circumstance that can be selected for.But unless you're proposing that cells actually determine the mutational outcome in response to need, what you're proposing is not at all different than a random mutation.Consider a casino. The fact that you can only gamble at card tables, and not the middle of the men's room, doesn't make the games themselves any less random. The way they correct them is by comparing sequences to their opposing side of the helix. They don't, for instance, predict the consequences of the mutation and then excise sequences that are deleterious. All genetic sequences are the same to the cellular repair mechanism, except insofar as they're complimentary to their opposite helix.The only process that removes mutations because of their survival consequences is natural selection. As I said, there is absolutely no difference between these situations. Both cases would be random mutation because the outcome of the mutation was not determined by need.It's certainly the case that some organisms respond to environmental stress by "turning up" the mutation rate of their DNA. This is still random mutation.This message has been edited by crashfrog, 01-23-2005 11:01 AM

Genes have, at most, one of two functions. They either regulate the expression of other genes; or they encode amino acid sequences for protein expression. Or they may have no function at all.It's proteins you're thinking of whose functions may change due to mutation. Proteins only get expressed in one way; the transcription mechanism you're probably already familiar with. And because we can read the genetic code we can read which sequences will generate proteins and which have "stop" codons right in the middle.

It may have lost a function, I guess, is that a change?If a gene that codes for a protein that (say) metabolizes chemical A mutates, and the resulting protein now metabolizes chemical B, the function of the gene hasn't really changed; it still just codes for proteins. What changed, to my mind, was the function of the protein.It's common to conflate, in conversation or explanation, the function of genes and their resulting products; I think it's important to at least consider the fact that the function is all in the protein.For genes that make proteins, anyway. That's a very good point. I thought I had covered all my bases but I see that I had not. Maybe I'll just bow out now and let the smart folks talk about this now.

By what mechanism? If random mutation is still happening, then you have evolution. I don't understand how your model is fundamentally different than the evolutionary one. At best you have a sort of add-on theory that posits additional evolutionary change. Maybe you could explain how your model prevents macroevolution from occuring, because that would be the inevitable outcome of truly random mutations.

But since we both agree this doesn't happen, I'm not sure what you're arguing. Well, we detect the formation of certain traits through mutation that there are no selection pressures for. For instance:

quote:

---

A reinterpretation by Kubitschek (1974) of work by Novick and Szilard (1956) suggests that the argument above was reasonable. In this study resistance to a bacterial virus was used as a marker to follow the appearance of some mutations in a chemostat culture. Novick and Szilard grew E. coli in a chemostat at a steady-state density of about 3 108 cells per ml. Periodically they assayed cells sampled from the chemostat for resistance to infection by bacteriophage T5 and calculated the density of T5 resistant cells in the culture. At no time was phage T5 present in the chemostat nor had the cells in the chemostat been exposed to phage T5. They found that there was always a fraction of cells in the culture that was resistant to T5. The density of resistant cells fluctuated betweeen 102 and 103 per ml. It followed a pattern like the one drawn below:

---

2,000 per ml
|                                                        *
R   |                                                       *
e   |                                                      *
s   |                             *                       *
i   |                            * *                     *
s   |                       *   *   *                   *
t   |                      * * *     *                 *
a   |                     *   *       *     * * * * * *
n   |                    *             *   *
t   |          *        *               * *
|     *   * *      *                 *
c   |    * * *   * *  *
e   |   *   *     *  *
l   |  *
l   | *
s   |___________________________________________________________
0
0                  Generations                             700

quote:

---

The increases and decreases reflect the occurrance of mutations within strains in the chemostat. The initial increase in the frequency of resistant cells occurs because a mutation occurs within a T5 resistant strain that makes it (and its descendents) the fastest growing cells in the culture. As long as this strain remains the fastest growing one its representation in the population will increase. Eventually different favorable mutation occurs in a cell that is sensitive to T5 that makes it (and its descendents) the fastest growing cells in the culture. This causes the frequency of T5 resistance to decline. Later a different mutation occurs in a T5 resistant strain that makes it the fastest growing strain. Its frequency increases, and so on.

​

It is important to note here that in this environment sensitivity and resistance to infection by T5 is a neutral trait here. Because there is no T5 in the environment, resistance does not provide an advantage. But it doesn't seem to provide much disadvantage either. If it provided a disadvantage, the resistant cells would washout of the chemostat. In this environment, it is selectively neutral. Mutations in other genes cause some cells to have a higher growth rate. It is just a matter of whether these mutations occur first in resistant or sensitive cells that determines whether the frequency of T5 resistant cells increases or decreases. It's a hitchhiking effect - the T5 resistance gene just goes along for the ride with the genes causing the fluctuations.

---

If you're at all math-minded, you should recognize that figure as a "random walk". And remember too that this is a chemostat monoculture, so there was no variation present in the inital population (bacteria are haploid so an individual has at most one allele per gene).We know that these mutations are occuring randomly and not in response to circumstance, because they occur in a random distribution, and occur in all circumstances. But we know that isn't true. Your model needs to work with the reality that some random mutations are, in fact, beneficial. Otherwise all you have is an ad-hoc rationalization of all beneficial mutations being created by your unknown mutagenic-on-purpose mechanism, without any actual evidence that this is so.We know that some random mutations can be beneficial. You can't simply ignore this. It's not as hard as you think. We classify species into "clades" based on their heredity. Family trees, if you will. No matter how much you change, you're always related to your relatives. So classification isn't that hard, thanks to genetics. (I say "not that hard" but it's actually a pretty involved procedure, one I'm lucky enough to be somewhat familiar with, due to my wife currently engaged in a line of research to establish the phylogenetics of several local species of an insect corn pest.) I don't know, exactly, what you mean by "stay the same." All individuals are variants, and populations are always changing. Nothing "stays the same" in biology; change and diversity are the rule. And what process, at the DNA level, is going to be able to predict which changes are going to be beneficial and which are not?Pop quiz: which is the beneficial mutation? Lots of fur and a squat, short body type, or no fur and a long, svelte form?Kind of depends on whether or not you live in Antarctica or the Sahara desert, doesn't it? Since the benefit of a mutation is inextricably linked to environment, natural selection is the only thing that determines whether or not a mutation is beneficial. At the DNA level, all mutations are equivalent.I simply don't see how your model can work, and it's based ont he faulty premise that you can simply ignore the fact that random mutations are occasionally beneficial. In fact, you can't determine the fitness outcome of a mutation short of its interaction with the environment, which your model doesn't even begin to take into account.

There's no difference, though. I think maybe you don't quite understand what we mean by "random mutation." "Random" doesn't refer to purpose, it refers to the outcome of the mutation, which is unpredictable. And of course "mutation" simply refers to genetic changes that occur as a result of imperfect copying or repair.An "on-purpose mistake" would still be a random mutation, because the change would be a mutation, and its outcome would be random.

I don't think he's proposing that; I think he's saying that because the mutation occured "on-purpose", that is a mechanism allowed or stimulated a mutation (but not the outcome of the mutation), the mutation was not "random".I disagree because that's not why we call them random. We call them random because the outcome of the mutation event is not deterministic; in that sense there's no difference between what he proposes and what we term "random mutation."

If it's not possible, then what's the difference?It may not be possible to tell the difference between gravity and another, previously unknown form of energy that I propose is what holds me, specifically (and not you), down onto my seat; but if it's not, what's the difference? Random mutation is somewhat controlled, by the cellular machinery that alread exists to repair genetic damage. Again, no difference between what you're proposing and what we already know provides the source of all genetic variation. The changes you're talking about aren't built into the cell, either. Under your proposal, the mechanism of the actual change is still copying errors or exterior mutagens. Your proposal is that cellular mechanisms exist that allow these exterior factors to exert greater effect; in a sense, that the DNA is left "exposed" to these mutations to a greater extent than normal as a specific cell response.What we're telling you is that this is already known to occur, and that its called "random mutation."The changes you're talking about in your model are still caused by copying errors and external mutagens, or else they would be deterministic changes, which you say that they are not. How else would a cell generate a random outcome? Again, we know this to be false. Random mutations have mostly no effect. Few are deleterious, and even fewer still are beneficial. Which is why mutation, for the large part, is restricted by cellular machinery. I think what you should be surprised about is how your model is exactly like the model you're trying to pit it against. You're quite on-base with your model; it's exactly what we find in nature, which is called "random mutation."

Purpose has nothing to do with mutation. Mutations are random not in that they're without purpose, but that the outcome of a mutation "event" is not determined by environment.Do you see, yet? We say that the lottery is random, because the outcome is not determined. (That would be fraud.) Yet, people don't play the lottery "randomly", or without purpose; they play for the purpose of gettin' paid.That doesn't make the lottery less random. You're conflating two meanings of the word "random": "nondetermined" and "without purpose." Again, no difference from random mutation as observed in nature, where certain genetic loci are known to mutate more frequently than others, though no segment of DNA is able to be completely shielded from mutation. Wait, now, you're backtracking. To variate not in certain ways means that some genetic changes are somehow disallowed; or that certain segments of DNA are not allowed to change.But you already agreed that you're not proposing this in your model. Which is it, exactly?