Info

Generation

Figure 12.7. Eight runs of computer simulation of idealized population of size 100 with fitness model described in Figure 12.3, all starting with initial condition of p = 0.3, and part of subdivided global population with overall frequency of 0.5 for A allele with subdivision corresponding to island model with Nm = 1.

shown in Figure 12.5 except for a global allele frequency of 0.3 instead of 0.1. In this case, four of the eight subpopulations have gone to the higher peak, and another is well on the way! This is the result of the shift in the balance of gene flow to selection and drift, with gene flow now favoring peak shifts by placing populations into the fitness valley where drift is most effective. Thus, even more demes go toward the higher peak, and the global allele frequency increases even more. Figure 12.7 shows the simulated results obtained when the global allele frequency has reached 0.5. At this point, gene flow has shifted toward a bias favoring the higher peak. It is now virtually inevitable for all the local demes to evolve toward the higher peak (Figure 12.7).

Wright argued that gene flow might be even a more powerful force in bringing more and more demes to a superior adaptive peak when there is an interaction between selection and the amount of gene flow. His argument applies to the case in which the superior adaptive peak results in a higher absolute fitness (recall that the adaptive surface is defined from relative fitnesses, so this is not necessarily the case). For example, in the j-Hb A, S, and C example, the superior adaptive peak is associated with fixation for C (Figure 11.8), and populations on this peak are characterized by 100% of the individuals having resistance to malaria that is superior to the resistance of A/S individuals, who represent only a minority of the individuals in populations on the lower peak. Hence, the absolute average viability of individuals (their ability to survive in a malarial environment) is indeed superior for those populations on the higher peak in this example. Wright felt that such a situation would often be true, so that those demes on the higher peaks would have more offspring than demes on lower peaks. As a result, the same migration rate per offspring would result in those demes on the higher peaks producing a greater number of migrants going out to other demes. This increase in the absolute numbers of migrants biases the evolutionary process to favor the spread of those alleles associated with the higher peaks throughout the species.

Wright (1931, 1932) called the above model of adaptive evolution the shifting balance theory, in which shifting balances between the relative strengths of selection, drift, and gene flow allow local demes in a subdivided population to explore the adaptive surface, then preferentially evolve toward the higher peaks in this surface, and ultimately draw other demes toward the higher peaks via asymmetric gene flow. From the onset, the shifting balance theory has been controversial (Coyne et al. 2000; Goodnight and Wade 2000; Whitlock and Phillips 2000). Nevertheless, shifting balance was not empirically tested until 60 years after it was proposed. This long gap between hypothesis and experiment stems in large part from the difficulty and labor involved in testing the shifting balance theory. Such tests require multiple coexisting demes and an overall large population size and manipulation and/or monitoring of many evolutionary parameters such as gene flow, drift, and selection. Wade and Goodnight (1991) provided these first empirical tests using laboratory populations of the flour beetle Tribolium castaneum and demonstrated that shifting balance can work under the appropriate circumstances. However, this does not resolve the question of how often the appropriate circumstances occur. Even though shifting balance is possible, it may play little role in adaptive evolution if those circumstances rarely occur. We will therefore examine some of the critical requirements of shifting balance to gauge how likely this mode of adaptive evolution may be.

Was this article helpful?

0 0

Post a comment