The effect of linkage on genetic variances within biparental simulated and Zea mays populations

Hybrid yield improvement in maize (Zea mays L.) has been attributed in part to the development of elite inbred lines within distinct germplasm groups adapted to distinct environments. Breeding largely within groups has caused the accumulation of distinct favorable alleles for adaptation, yield, and yield component traits. Favorable alleles may be linked to other favorable alleles in coupling phase or linked to unfavorable alleles in repulsion phase. To investigate the effects of linkage on genetic variances, we used a simple, two-locus model to simulate populations of intermated and standard recombinant inbred lines (RILs) derived from parental lines with coupling or repulsion phase loci. Genetic variances differ between simulated intermated RILs (IRIL) and RIL testcross populations as a function of population size, the effect sizes of linked loci, and the genetic distance between loci. We also generated RIL and IRIL testcross (TC) populations from two short-season inbred lines from different heterotic groups and evaluated yield, three yield component traits, and five flowering time–associated traits. For all nine traits, the inbred lines have accumulated distinct favorable alleles. Genetic variances are high and transgressive phenotypes are frequent in both testcross populations. Intermating does not increase genetic variation. Genetic variances of six traits within the RIL-TC population are nominally between 1.25 and 1.75-fold greater than variances within the IRIL-TC population. We conclude that the genomes of the parental inbred lines may harbor coupling rather than repulsion phase loci and heterosis is unlikely due to pseudo-overdominance.