Relative effect of drift and selection in diverging populations within a reciprocal recurrent selection program

The establishment of a solid heterotic pattern has been beneficial to maize (Zea mays L.) breeding in North America. It has been shown that genetic divergence between heterotic groups has increased over time, but the relative importance of drift versus selection for hybrid performance on the increased divergence is not known. Reciprocal recurrent selection is a systematic method of selection on hybrid performance, where two populations are selected and improved on the basis of the population-cross performance. The University of Nebraska performed a replicated recurrent selection program (UNL-RpRS) for eight cycles of selection. Three replicate programs were conducted in parallel for both reciprocal full-sib selection (RFS) and S1-progeny selection. The objective of this study was to evaluate the effect of selection for hybrid performance on population divergence using the UNL-RpRS as a model system. Genome-wide divergence among the base populations increased over eight cycles of selection. Divergence was not, however, greater among base populations subjected to RFS compared with S1-progeny selection. Moreover, Wright's fixation index (F_st) values among replicate populations within a base population and selection method were as large as F_st values between complementary populations selected for population cross performance using RFS. A small increase in divergence through RFS relative to S1-progeny selection was observed when markers were filtered for high levels of adjacent linkage disequilibrium (LD), but the effect was consistent across replicates. We did not consistently detect effects of selection for hybrid performance on genome-wide divergence between populations, suggesting selection was not a strong force in diverging populations in the UNL-RpRS.a