next up previous
Next: Two-locus two-allele model Up: Deterministic models Previous: Genetic assumptions

The rate of evolution, general treatment

There has been substantial progress in the mathematical theory of multi-locus systems in recent years (Nagylaki, 1993; Barton and Turelli, 1987, 1991; Bürger, 1993). One of the main results is that the dynamics of the variance critically depends on genetic details. Therefore the present paper analyzes a number of explicit genetic models. In this section we will derive a continuous time version of the general two-character equations for discrete generations of Turelli (1988a). This version is specialized to the features of the corridor model. The main aim is to aid the analysis of the special genetic models. It will be shown that the consequences of selection on deleterious pleiotropic effects are reflected in the dynamics of the genetic variance of the character under stabilizing selection.

Let be the genotypic value of the genotype comprised of gametes i and j, and let be the average effect of a gamete i on character , and finally let be the frequency of this gamete. Then, under random mating, the mean genotypic value of character 1, , is

With additive effects, the genotypic values are simply the sum of the gametic effects. Consequently, the mean genotypic value is two times the mean gametic effect where is a gametic effect on character 1. Hence, the rate of change of the character mean of is twice the rate of change of the mean gametic effect

where the rate of change in the gamete frequencies is given by

In this equation is the mean fitness of the gamete i, the mean fitness of the population and the increase or decrease of gamete frequency caused by recombination. The rate of evolution then becomes

With additive effects the second term on the right hand side is always zero, because the re-distribution of alleles caused by recombination alone will not change the mean value of the character. Expanding the first term leads to

This is the 'seconday theorem' of Robertson (1966). The above derivation shows that it is exact for additive characters. Applying this formula to the fitness function of the corridor model, one obtains

An analogous formula holds for discrete generations. The first term on the right hand side is the classical result about the response to directional natural selection (Lande, 1976). The second term concerns the interaction between evolution of the first character and fitness effects of the second. It corresponds to the term in Turelli's two character equation (Turelli, 1988a). Note that this term is not the correlated selection response, as will be shown below. The precise meaning of this term can only be explicated in the context of an explicitly genetic model. It will be shown below, that the term describes the effect of stabilizing selection on the genetic variance of the second character. This variance is linked via pleiotropic effects to the evolution of the first character.



next up previous
Next: Two-locus two-allele model Up: Deterministic models Previous: Genetic assumptions

Tue Apr 9 13:43:34 EDT 1996