Genomic prediction of forage nutritive value in perennial ryegrass

Abstract

Background

Despite its importance to animal production potential, genetic gain for forage nutritive value has been limited in perennial ryegrass (Lolium perenne L.) breeding. The objective of this study was to phenotype a training population and develop prediction models to assess the potential of predicting organic matter digestibility (OMD) and neutral detergent fiber (NDF) with genotyping-by-sequencing data.

Methods

Near infra-red reflectance spectroscopy calibrations for OMD and NDF were developed and used to phenotype a spaced plant training population of n = 1606, with matching genotype-by-sequencing data, for developing genomic selection models. $F_2$ families derived from the training population were also evaluated for OMD and NDF in sward plots and used to empirically validate prediction models.

Results

Sufficient genotypic variation exists in breeding populations to improve forage nutritive value, and spectral bands contributing to calibrations were identified. OMD and NDF can be predicted from genomic data with moderate accuracy (predictive ability in the range of 0.51–0.59 and 0.33–0.57, respectively) and models developed on individual plants outperform those developed from family means. Encouragingly, genomic prediction models developed on parental plants can predict OMD in subsequent generations grown as competitive swards.

Conclusions

These findings suggest that genetic improvement in forage nutritive value can be accelerated through the application of genomic prediction models.