Evidence against Fisher's theory of dominance

Abstract

FISHER'S theory of the evolution of dominance is based on the fact that heterozygotes for rare deleterious alleles, maintained by recurrent mutation from the wild-type allele at the locus, greatly outnumber homozygotes in a random-mating population at equilibrium¹. The fitness of a heterozygote carrying a mutant and the wild-type allele can be written as 1 − hs (1 ⩾ h ⩾ 0), compared with fitnesses of 1 and 1 − s for wild-type and mutant homozygotes respectively. Fisher postulated that mutant alleles were originally semidominant, so that h > 0. He explained the fact that most mutations are observed to be recessive or nearly recessive to wild type as the result of selection for modifier genes which increase the fitness of the heterozygotes alone. He pointed out that selection to improve the fitness of the mutant homozygotes is likely to be relatively ineffective because of their rarity compared with the heterozygotes, and also because the presence of the wild-type allele in heterozygotes renders them more amenable to selection towards the wild type. Using this theory, the level of dominance of the mutant alleles, h, is gradually reduced towards zero, while s remains constant. Wright² criticised this because the heterozygotes for mutant alleles are so rare that the intensity of selection on a gene which modifies h is of the same order as the rate of mutation, u, to deleterious alleles at the locus undergoing dominance modification. He suggested that such a weak selection pressure is unlikely to overcome the effects of random genetic drift, or the selective consequences of pleiotropic effects of the dominance modifier. Later theoretical work^3,4 has confirmed Wright's conclusion about the intensity of selection on dominance modifiers, so that many geneticists have come to doubt the validity of Fisher's theory³. The principle of dominance modification when applied to balanced polymorphisms (for example, Batesian mimicry⁵) is not, of course, subject to this objection and is generally accepted. Fisher¹, however, believed that the recessivity of mutant genes demonstrated the ability of even a minute selection pressure to cause significant evolutionary change. The issue cannot be settled by purely theoretical considerations, and it is evident that some empirical evidence capable of providing a critical test is required. (Demonstrations that dominance can be modified by artificial selection^6,7, although of great interest, are not relevant to the question of the efficacy of selection pressures as small as those postulated in Fisher's theory.) This paper points out recent experimental evidence on the heterozygous effects of mutations affecting viability in Drosophila⁸ that seems to be inconsistent with Fisher's theory, and also looks at two alternative theories proposed by Haldane⁹.