Natural Limitation to Evolutionary Processes (2/14/05)

quote:

---

[About genetic drift]
Funny it's categorized on so many biology sites as a "process of evolution" then. I myself am pointing out that all those processes (with the exception of immigration and mutation) are processes of selection. For heaven's sake it's obvious. It selected out the cheetah.

---

There is a problem here with communication, caused by your lack of familiarity with scientific definitions. "Selection" means natural selection in biology, and applies only to alleles with positive or negative effects. Neutral alleles are not selected by drift or by anything else, by definition. (You had a similar confusion earlier when you said that PBS meant "mutation" when it said evolution. No, it meant evolution, which as defined by science includes frequency changes in existing alleles.)On one point you are correct: genetic drift does reduce genetic variation. Depending on how you choose to measure variation, drift either always decreases variation or both increases and decreases it.As far as I can tell, however, that is one of the few true statements you have made about genetics here. Scientists actually do know quite a lot about mutation and drift, and they are not all extremely stupid, which they would have to be not to have noticed that evolution doesn't make any sense. We can measure the mutation rate in a wide variety of organisms. We can calculate how much variation there should be in a given size population, based on the mutation rate (and assuming most mutations are neutral). We can measure the amount of variation actually present in populations, and when we do we find that everything makes good sense. (Not perfect sense -- the calculations are approximations, and there are plenty of details to worry about, but the big picture is clear enough.)Let me be concrete. Based on the measured mutation rate, it is likely that every single base in the human genome has mutated at least once just in the current generation. That represents an enormous input of variation, in just a single generation. Much of that variation will disappear immediately: roughly a quarter of those mutations will disappear (by drift) in the next generation, and more in the generation after that. But each of those generations will introduce its own large load of new variation. In smallish population, the forces increasing and decreasing variation are often in balance, so that the total amount of variation stays about the same. Humans, as it happens, have too large a population, and too long a generation time, for the forces ever to come into balance, barring a massive population collapse. So for us, genetic variation is actually increasing.

quote:

---

Excuse me??? It sounds like YOU are denying that a genetically stable population happens at all!!

---

It depends on what you mean by genetically stable. There are populations in which the total amount of variation stays about the same. There are no populations in which the actual variants present in the population stay the same over any significant length of time.

quote:

---

I may not be following you at this point, but my observation is that the processes that SELECT the variations REDUCE genetic variability and that this is the overall tendency in ALL processes of speciation or "evolution."

---

This is true, more or less. What it neglects is the process of mutation, which is constantly adding more variation to the population, both during speciation and afterwards. Suppose a small subpopulation get separated from the larger group, perhaps in a different environment. Natural selection in the new environment and drift in the small group will indeed reduce genetic variation, and may lead to a new species. If the new species is successful in this environment, however, its population will expand, and it will immediately start accumulating more genetic variation. Small populations have little variation because there simply aren't enough individuals to carry many variants. With more individuals, more of the new mutations (which occur every generation) will persist in the population.

quote:

---

I believe that much of what is called "mutation" is in fact merely pre-existing allelic potentials that come to the fore as combinations typical of the parent population are eliminated from the new population.

---

Beliefs are better when they correspond to reality; your belief here doesn't. "Mutation" is defined as a change in genetic material; if it already existed, it isn't a mutation. When biologists study mutation, they're studying changes to genetic material, and it is those mutations that we are talking about. It's already been explained to you how you can observe new mutations occurring in bacteria. Perhaps you should go back and think about that example some more, and see if you can formulate a response to it (other than simply dismissing it).

quote:

---

Or do you all think it is brought about by mutations in the sense of NON-pre-existing accidents

---

Yes. As far as we can tell, all genetic variation comes from mutations, in the sense of non-pre-existing accidents.

quote:

---

If so, somebody should tell all the biology sites that say evolution is change in gene or allele frequencies over time

---

Why? Evolution is change in allele frequencies over time. Mutation introduces new alleles, and the processes of evolution change the frequencies of those alleles. It seems pretty straightforward to me. What's wrong with it?

quote:

---

You're even wrong to claim that the Hardy-Weinberg equation always points to declines in variation - it can also point to equilibrium situations (as it does in the case of sickle-cell anaemia).

---

Why do Hardy and Weinberg get the credit (or the blame) for all of population genetics? Hardy-Weinberg describes the relationship between allele frequencies and genotype frequencies in the case of an infinite population, random mating and no selection, nothing more. In Hardy-Weinberg equilibrium, variation does not decline (nor does it increase), nor do allele frequencies change, in response to selection or anything else. No real population achieves Hardy-Weinberg equilibrium, of course, but it is a useful approximation for many purposes.

quote:

---

But here is the important part, that I fear you are missing: Even though the success of the mutated allele drops "diversity" for its gene to zero, the organisms within the population have NOT reduced the allelic diversity for the other 20,000 genes in their genome. Thus, fixation of an allele for one gene generally has neglible effects on the allelic diversity for the other 99.99% of the genome. Genes coding for any other trait the population have retained their allelic diversity.

---

Here's an example from close to home. People with ancestry from northern Europe are a little unusual in that they are probably able to digest milk as adults; for most people, lactose intolerance sets in after childhood. Lactose tolerance in Europeans is the result of natural selection acting on a single genetic variant on chromosome 2 sometime in the last five or ten thousand years. Selection did substantially reduce genetic variation around the gene in question (the gene for lactase), in a region about 1 million bases across. The remaining 99.97% of the genome is largely unaffected. Note that this was the result of a powerful selective pressure; weaker selection would have produced a smaller affected area. (Reference: Genetic Signatures of Strong Recent Positive Selection at the Lactase Gene, Bersaglieri et al, Am. J. Hum. Genet., 74:1111-1120, 2004.)

quote:

---

Isn't that the field known more commonly as "bioinformatics"?

---

More like population genetics.

quote:

---

The only question is whether mutation increases genetic diversity enough to keep the theory of evolution afloat.

---

Your thinking on this entire issue is backwards. The question isn't whether mutation can increase genetic diversity enough to counteract the other forces acting in evolution. It isn't the question because the rate of the other forces depends on how much mutation is occurring. If there is a very low mutation rate in some organism, then there will be very little new variation introduced. Little variation means that selection will act only very rarely, and that little drift will occur. Bottlenecks in such an organism's population will remove little diversity, because there wasn't much there to start with. It makes no sense to ask whether in theory mutation can keep up with these processes, since it's mutation that drives them.What would be a meaningful question is this: is the amount of genetic variation we see consistent with what we would expect if mutation were the only source of variation, and given what we know about genetic drift (which is a great deal) and about selection (not so great). Or is there too much variation, which might be the case if large populations of organisms had recently been created de novo, or too little variation, which would be the case if the Flood was a historical event. This is an experimental question, and the answer is that the amount of variation is indeed just about what we would expect if it all comes from mutation.

quote:

---

Isn't the Hardy-Weinberg Equilibrium equation often used as an indicator that evolution is happening? You can compare allelic frequencies from the equation to actual allelic frequencies in the population to show that/if evolution is happening.

---

I've seen HW equilibrium talked about as a test of evolution, but it is very rarely a useful one; it's just not sensitive enough, nor is it specific as to the cause of disequilibrium, even if you can detect it. The idea is that you can test whether individuals with two copies of an allele are favored by testing whether there are more of them in the population than you would expect from the allele frequency. This will work in the case of strong balancing selection, in which case you will find more heterozygotes (one copy of each allele) than you expect; I think you can detect selection acting on the sickle cell allele this way. (One copy of the allele protects you from malaria, but two copies gives you sickle cell anemia.)In general, however, you need to look at very large sample sizes in order to detect the slight shifts in genotype frequencies that you get even in pretty strong selection. And once you have sample sizes that large, you're more likely to be detecting subtle substructure in the population or assortative mating than you are selection.HW equilibrium testing is widely used in human genetics for something else, however: it's a test of genotyping accuracy, i.e. it's a check for experimental error. If everyone in your sample tests as having one copy of each allele, there's something wrong with your test.

quote:

---

Here is my scenario:
I have a population of 1 million individuals. This is the carrying capacity of the available resources.

​

These animals live 5 years producing two surviving pups per female per year from the end of their first year.

​

They have 30,000 genes. There is no junk DNA. There are therefore 30,000 alleles in the whole population.

​

Each individual has 10 mutations randomly scattered in it's genes.

​

Half of these are fatal and the pups carrying them are still born or not even carried to term. This gives 5 mutations per individual.
[...]
Now what I need to do is figure out what the equilibrium diversity will be.

---

This isn't exactly your scenario, but it's pretty close:Population of 1 million, with a generation time of 5 years. Each female produces a random (i.e. Poisson) number of offspring that live to reproduce, with a mean of two. (Treating the number as Poisson is a simplicification of your scenario, in which the number is capped at ten and has some granularity.) The effective mutation rate is 5 per diploid genome per generation, or 2.5 per haploid genome.For this scenario, and if we assume a very large genome, such that we can ignore multiple mutations occurring in the same base, then at equilibrium the mean number of differences between any two chromosomes is 4Nu, where N is the (effective) population size and u is the mutation rate. For your parameters this is ten million, i.e. if you compare two random chromosomes from the population they will have ten million differences. The total number of variants in the entire population is (on average) 4Nu x sum(1/i), where the sum runs from 1 to N-1; this is approximately 4Nu(ln(N) + 0.577) for large N, or 144 million.You won't get to equilibrium any time soon, however. The characteristic time here is 2N, so you'll have to wait a few tens of millions of years to be near equilibrium.(In reality, effective population sizes are usually a good deal smaller than a million, at least for the kind of organisms I'm used to dealing with. 10,000 to 100,000 is typical.)