Description
Alfalfa (Medicago sativa L.) is a perennial outcrossing legume that is cultivated as an important forage crop in many parts of the world. Yield is the most important trait for profitable alfalfa production, yet over the last 30 years yield improvement in California has stagnated. Current breeding methods focus on recurrent phenotypic selection; however, alternatives incorporating genomic- and phenomic-based information may enhance genetic gain and help to address the lack of yield improvement. Here we attempt to increase the yield potential of alfalfa using genomic selection (GS) in combination with high throughput phenotyping (HTP). A total of 193 families from two closely related elite populations were sown in the greenhouse and transplanted into mini sward plots at two locations near Davis, CA in May 2020. The trial was managed as a high-input system under full irrigation. Families were genotyped and phenotyped for biomass yield by mechanical harvest and a combination of drone and tower-based remote sensors across 12 harvests, 3 in the establishment year (2020), 7 in the first full year of production (2021) and 2 in 2022. Alfalfa yields ranged from 13-27 tonnes DM/hectare/year with a number of half-sib families outperforming popular cultivars in the first 2 years of production. Biomass volume predicted from the drone-based cameras had a moderate prediction accuracy with an overall R2 of 0.55. Some individual harvests reached accuracies as high as 0.85. Genotyping resulted in a dataset with 6,838 SNPs. Allele frequencies were used to generate a relationship matrix for GS. Narrow-sense heritability for dry matter yield was 0.31 and the predictive ability of the GS model was 0.15.
Included in
Agricultural Science Commons, Agronomy and Crop Sciences Commons, Plant Biology Commons, Plant Pathology Commons, Soil Science Commons, Weed Science Commons
Prospects for Improving Alfalfa Yield Using Genomic- and Phenomic-Based Breeding
Alfalfa (Medicago sativa L.) is a perennial outcrossing legume that is cultivated as an important forage crop in many parts of the world. Yield is the most important trait for profitable alfalfa production, yet over the last 30 years yield improvement in California has stagnated. Current breeding methods focus on recurrent phenotypic selection; however, alternatives incorporating genomic- and phenomic-based information may enhance genetic gain and help to address the lack of yield improvement. Here we attempt to increase the yield potential of alfalfa using genomic selection (GS) in combination with high throughput phenotyping (HTP). A total of 193 families from two closely related elite populations were sown in the greenhouse and transplanted into mini sward plots at two locations near Davis, CA in May 2020. The trial was managed as a high-input system under full irrigation. Families were genotyped and phenotyped for biomass yield by mechanical harvest and a combination of drone and tower-based remote sensors across 12 harvests, 3 in the establishment year (2020), 7 in the first full year of production (2021) and 2 in 2022. Alfalfa yields ranged from 13-27 tonnes DM/hectare/year with a number of half-sib families outperforming popular cultivars in the first 2 years of production. Biomass volume predicted from the drone-based cameras had a moderate prediction accuracy with an overall R2 of 0.55. Some individual harvests reached accuracies as high as 0.85. Genotyping resulted in a dataset with 6,838 SNPs. Allele frequencies were used to generate a relationship matrix for GS. Narrow-sense heritability for dry matter yield was 0.31 and the predictive ability of the GS model was 0.15.
