Implementation of rapid cycle recurrent genomic selection for forage yield in perennial ryegrass

Abstract

Opportunities exist to accelerate genetic gain in forage breeding using genome-wide selection approaches. In this study, we evaluated rapid cycle recurrent genomic selection (GS) as a means of improving genetic gain for value of annual forage yield. A small population of tetraploid half-sib families was evaluated for seasonal forage yield over two years, and the maternal parent plants were genotyped and genomic prediction models developed. The GS model for value of annual forage yield had a predictive ability of 0.23. An initial round of among-family selection based on field evaluations and within-family selection using genomic estimated breeding values was performed. This was followed by two further GS cycles. New synthetics were produced after each round of selection and were established in a field trial alongside the starting population. A positive response to selection was observed in new synthetics after two successive rounds of rapid cycle recurrent genomic selection before declining in the third round. The genetic gain for the value of annual forage yield was 2.4% from C0 SYN-1 to C1 SYN-1 and 6.4% from C1 SYN-1 to C2 SYN-1. In the case of C0 to C1, genetic gain was compounded by among-family selection based on field evaluations. The implementation of rapid cycle recurrent genomic selection offers an opportunity to increase genetic gain; however, the predictive ability is likely to decay rapidly as selection candidates become more distant from the training population.

Keywords

• forage yield
• genomic selection
• lolium perenne
• perennial ryegrass