THE EFFICIENCY OF USING DOUBLED HAPLOIDS IN A RECURRENT SELECTION PROGRAM IN A DIPLOID, CROSS-FERTILIZED SPECIES

Abstract

Mass selection in both doubled haploid and diploid populations was simulated for a 30-year period under additive and complete dominance models. The response to S1 selection was also included for the purpose of comparison. All selection programs were conducted at two selection intensities, 5% and 25%, under 0.2 heritability with a constant population size of 400. S1 lines were evaluated in 10-plant plots with four replications; use of a winter nursery was assumed. The character under selection was controlled by 20 major (A = 2) genes and 20 minor (A = 1) genes. The initial frequency of ten of the major genes and ten of the minor genes was 0.5; that of the remaining ten of each type was 0.1. The response to mass selection with doubled haploids was about 1.4 times faster than diploid mass selection and equal to S1 selection in terms of gain per year. If replicated trials were used to evaluate the doubled haploids, the rate of gain was faster than in S1 selection. The greater efficiency of using doubled haploids was due to doubling of the additive genetic variance and the elimination of dominance variance. However, significant loss of desirable genes occurred in the mass selected doubled haploid population under both the 5% and 25% selection intensities in the additive and complete dominance models.