Replies

What you say is literally true, but irrelevant. As Edsheppa pointed out, what we are really talking about (and what we should be talking about) is the probability that a mutation exists in a population at a point in time. To begin to calculate that, we have to ask ourselves what the probability is that a mutation would occur. The second question is what the percentage is of individuals with the mutation over the next relevant time span. I contend that, assuming no natural selection, these 2 numbers are identical or just about. I think Edsheppa disagrees with the latter proposition.

Actually I was responding to a different point you made a while back - something about mutations dieing out.

Just to be clear here, the one is the probability that the mutation occurs in any one reproductive event and the other is the long term percentage of instances of the mutation in the population. I do think they're different because the mutation is heritable so percentages of its occurrence in successive generations are correlated - you're assuming indepedence I think. A manifestation of this correlation is that once the mutations appears is has a probability of becoming fixed that's basically the number of occurrences divided by the population size. Thereafter it remains prevalent until the un-mutation (I assume its equally unlikely as the mutation) occurs and then becomes fixed. Obviously that could take a very long time.

A bigger point is you're making the same mistake as the article's author - the likelihood of one particular outcome (in this case a set of particular neutral mutations) is not very interesting. One must figure what the set of "interesting" outcomes are and compute that probability.

Well, I wasn't convinced by my reasoning so I wrote a simple simulation of an asexual population. The poulation is kept in a range. Normal individuals can mutate and mutants can umutate with the sampe probability. As seems obvious in retrospect, for this simple case in the long run an individual has a 50% chance of being a mutant.

To begin to calculate [the probability that a mutation exists in a population], we have to ask ourselves what the probability is that a mutation would occur. The second question is what the percentage is of individuals with the mutation over the next relevant time span.

Once a mutation occurs, the probability of it passing to the next generation or through a population is either random (via drift--in which case the probability moves toward 1 or 0) or follow any of a number of selection models. These models are based on the sorting of genes, or more inclusive sections of the genome called loci, within a population. The sorting of individual mutations follows the sorting of the loci they are located in.