When is early generation selection effective in self-pollinated crops?
Despite widespread use of early generation testing and selection (EGT) in breeding for self-pollinated crops, its effectiveness remains largely an unresolved issue. This issue is tackled here using elaborated genetic models that enable genetic and nongenetic effects to be assessed for the effectiveness of EGT in terms of (i) the selection response at one or more early generations relative to the response to direct selection at homozygosity and (ii) the probability of retaining superior lines selected during EGT. The selection response to EGT is analyzed for a model with quadratic genetic components due to additive, dominance, additive x additive, and linkage effects in selfed populations derived from a cross between two inbreds. The response to one cycle of EGT is less than the response to direct selection and decreases with nonadditive effects, repulsion linkage and reduced heritabilities. The cumulative response to two or more cycles of EGT is greater than the response to direct selection unless there are strong nonadditive effects, strong repulsion linkage, and low heritabilities. The probability of retaining a superior line at EGT decreases with increased nonadditive effects and low heritabilities. The proportion of lines needed to minimize the risk of erroneously culling superior lines increases with increased nonadditive effects and low heritabilities. Thus, EGT should be used for populations or traits with little nonadditive effect, coupling linkage and high heritability.