In article <3mjlut$sj6 at newshost.lanl.gov>,
Martijn A. Huynen <mah at primer.lanl.gov> wrote:
>>Does anyone know how the expected frequency of a specific nucleotide
>(or allele), and the distribution of this expected frequency
> depend on the mutation frequency in a case of neutral evolution.
>Kimura describes the change in
>the frequency-distribution in absence of mutations, but this (of course)
>allways goes to fixation. I need a formula that both includes mutations
>and genetic drift.
If the mutation model has the property that the probability of a
particular base arising is u, irrespective of what other base is there
(a property that is met by the Jukes-Cantor mutation model but byt few others)
and if the rate of mutation away from that base is v, then
the density function of allele frequency p in a population of
effective size N is
4Nu-1 4Nv-1
f(p) = C p (1-p)
where C is (I think) Gamma(4Nu+4Nv)/(Gamma(4Nu)Gamma(4Nv)) or at any
rate whatever is needed to get the integral between 0 and 1 to be 1.
The distribution is a beta-distribution. This is an old result of
Sewall Wright from 1930.
-----
Joe Felsenstein, Dept. of Genetics, Univ. of Washington, Seattle, WA 98195
Internet: joe at genetics.washington.edu (IP No. 128.95.12.41)
