Genetic architecture of reproductive fitness and its consequences

Reproductive fitness characters are crucial in animal and plant breeding, evolutionary genetics and conservation biology. However, the nature of their variation is not well understood. I review evidence on the comparative architecture of genetic variation for fitness and peripheral characters for both wild populations and domestic livestock. Fitness characters differ from peripheral characters in typically having lower heritabilities, directional dominance, higher levels of gene interactions, asymmetry of selection response, inbreeding depression and heterosis, declines in mean from mutation, and non-linear changes in genetic variation with inbreeding. These differences are a reflection of the different forces of natural selection operating on the two classes of traits. Genetic diversity for fitness in wild populations is partially due to rare, deleterious, partially recessive alleles in mutation-selection balance, with a further component due to alleles at equilibrium due to balancing selection, but there is no consensus about the relative contributions. In contrast, livestock populations are unlikely to be in equilibrium, due to changes in their environments, artificial selection, small effective population sizes and in some cases to crossing of populations. These have major effects on the genetic architecture of fitness and especially on the proportion of polymorphic loci exhibiting overdominance. In general, long-term directional artificial selection on peripheral traits is expected to move their genetic architecture towards that of fitness traits. Whilst the breeders' equation provides good predictions of selection response for peripheral characters, it does not predict asymmetrical responses to selection as observed for fitness characters, and levels of inbreeding depression and heterosis remain unpredictable. Given the importance of fitness characters and the uncertainties on many important issues, they deserve a higher priority in quantitative genetics research.