Optimizing Resource Allocation in a Genomic Breeding Program for Perennial Ryegrass to Balance Genetic Gain, Cost, and Inbreeding.

Genomic selection (GS) has been recognized as offering numerous potential benefits for ryegrass (Lolium perenne L.) breeding. While the theoretical benefits of GS in ryegrass breeding are clear, the best way to apply GS in practical breeding programs remains to be determined. The present study aimed to investigate genomic breeding options that best balance genetic gain, breeding costs, and the level of inbreeding using stochastic simulation. Nine GS scenarios were tested, including different numbers of selection candidates (10,000, 5000, and 2000 F1 seedlings) and three reference population sizes for GS composed of plots representing a sward-based trial (500, 200, and 100 plots). Low to moderate prediction accuracy was achieved for productivity traits across cycles (i.e., 0.1–0.45 for yield [h2 = 0.3]). Scenarios with larger reference populations (i.e., 500 plots) achieved higher prediction accuracy but, when considering field trial costs, were more expensive per unit of genetic gain. All nine GS scenarios delivered significantly higher genetic gain (up to fourfold) than the conventional breeding scenario over a 20-yr period. Scenarios with moderate selection intensity on F1 seedlings with fewer plots tested in field gave the most genetic gain per dollar invested (i.e., 2000 or 5000 F1 seedlings and 100 plots). However, all GS scenarios reduced genetic diversity in the breeding population more than phenotypic selection, highlighting the need to mitigate inbreeding when applying GS in perennial ryegrass. [ABSTRACT FROM AUTHOR]