Articles
Gene discovery and genomic prediction for improved strawberry flavor
Article number
1465_25
Pages
147 – 154
Language
English
Abstract
Flavor is a key trait underlying consumer acceptance of strawberry fruit.
Due its subjective nature and difficulty of evaluation at scale, flavor has not realized the same consistent improvement as objectively measured agronomic traits.
To improve flavor in the University of Florida strawberry breeding program, two complimentary approaches are taken, candidate gene discovery with molecular marker development for key aroma compounds and breeder-based sensory evaluations for genomic prediction.
Using 400 genetic individuals, a genome-wide association study (GWAS) was conducted for γ-lactones.
A novel locus was discovered associated with γ-dodecalactone production and a high-resolution melting marker was developed for implementation in marker-assisted seedling selection (MASS). Breeder-based sensory evaluations were carried out across 2 years and 16 total weeks on 800 genetic individuals following a sparse testing design.
Genotypes were evaluated for flavor, sweetness and sourness by 5 tasters.
Genomic prediction models were trained on various configurations of training populations and tested in 80:20 cross-fold validation or using trained sensory panel data using 10,000 SNP markers.
Correlation between predicted value and actual value was moderate for flavor (0.33-0.35), sweetness (0.33-0.34) and sourness (0.34-0.37). Using genotypes with trained sensory panel data as a testing population, correlations were moderate to strong for flavor (0.31-0.37), sweetness (0.41-0.49), sourness (0.66-0.70).
Due its subjective nature and difficulty of evaluation at scale, flavor has not realized the same consistent improvement as objectively measured agronomic traits.
To improve flavor in the University of Florida strawberry breeding program, two complimentary approaches are taken, candidate gene discovery with molecular marker development for key aroma compounds and breeder-based sensory evaluations for genomic prediction.
Using 400 genetic individuals, a genome-wide association study (GWAS) was conducted for γ-lactones.
A novel locus was discovered associated with γ-dodecalactone production and a high-resolution melting marker was developed for implementation in marker-assisted seedling selection (MASS). Breeder-based sensory evaluations were carried out across 2 years and 16 total weeks on 800 genetic individuals following a sparse testing design.
Genotypes were evaluated for flavor, sweetness and sourness by 5 tasters.
Genomic prediction models were trained on various configurations of training populations and tested in 80:20 cross-fold validation or using trained sensory panel data using 10,000 SNP markers.
Correlation between predicted value and actual value was moderate for flavor (0.33-0.35), sweetness (0.33-0.34) and sourness (0.34-0.37). Using genotypes with trained sensory panel data as a testing population, correlations were moderate to strong for flavor (0.31-0.37), sweetness (0.41-0.49), sourness (0.66-0.70).
Publication
Authors
M. Porter, X. Song, J. An, Z. Fan, Y. Wang, V. Whitaker, S. Lee
Keywords
strawberry, flavor, genomic prediction, GWAS
Groups involved
- Division Vine and Berry Fruits
- Working Group Strawberry Culture and Management
- Division Sustaining Horticulture in a Changing World
- Division Horticulture for Human Health
- Division Plant Genetic Resources, Breeding and Biotechnology
- Division Greenhouse and Indoor Production Horticulture
- Division Precision Horticulture and Engineering
Online Articles (64)