Understanding and using quantitative genetic variation

Abstract

Quantitative genetics, or the genetics of complex traits, is the study of those characters which are not affected by the action of just a few major genes. Its basis is in statistical models and methodology, albeit based on many strong assumptions. While these are formally unrealistic, methods work. Analyses using dense molecular markers are greatly increasing information about the architecture of these traits, but while some genes of large effect are found, even many dozens of genes do not explain all the variation. Hence, new methods of prediction of merit in breeding programmes are again based on essentially numerical methods, but incorporating genomic information. Long-term selection responses are revealed in laboratory selection experiments, and prospects for continued genetic improvement are high. There is extensive genetic variation in natural populations, but better estimates of covariances among multiple traits and their relation to fitness are needed. Methods based on summary statistics and predictions rather than at the individual gene level seem likely to prevail for some time yet.

Figure 1.

Changes in milk yields of US Holstein cows: phenotypic mean yields (P), mean breeding values (A) and environmental effects (E = A − P) derived from USDA data. Results are given relative to 1957, when the mean yield was 5859 kg. (Adapted from http://aipl.arsusda.gov/eval/summary/trend.cfm).

Figure 2.

Responses to selection for oil content in maize in the Illinois selection lines. Line designations: IHO (light yellow square), continued selection for high oil, ILO (dark yellow square), for low oil; RHO (green triangle), RLO (white circle), reverse selection; SHO (black square), re-reversed selection. (Adapted from Dudley & Lambert 2004).