Re: Reply to Rindos re genetic drift

Dave Rindos (arkeo4@UNIWA.UWA.EDU.AU)
Thu, 19 May 1994 06:45:07 +0800

On Wed, 18 May 1994, Read, Dwight ANTHRO wrote:

> Wilk writes:
>
> "Just a short note; I seem to remember from a research paper on genetic
> drift I wrote as a grad student, that Wright actually demonstrated
> mathematically that drift could work in quite large populations, as long
> there is little selective pressure on two alleles. The math was quite
> elegant and simple."
>
> I think Wilk is referring to the work of Kimura (and others) on neutral
> mutations where it is showed that an allele can be fixed in a population via
> a random walk process when there is little or not selection acting on the
> allele.

Yea... THE *ultimate* case of this phenomenon, using an example I have
given in classes, runs as follows:

Imagine two different alleles. They have *exactly* the same fitness value
within a population of any size (s, the fitness difference = 0). Matter
of fact, they are *identical* in all respects except that one is named
Pete and the other is named Bill. They start off at a specific, given,
frequence. Could Pete replace Bill in the population? [Of course!]

Now consider a realistic population size, and a finite amount of
time. What is the PROBABILITY that this stochastic replacement will
happen?

Who wants to bet on Pete winning and who wants to bet on Bill winning?

How much would *you* wager on this bet?

Dave,
contemplating yesterday's post about people and statistics...