Empirical evaluation of multi-trait DNA testing in an apple seedling population

Abstract

Marker-assisted seedling selection (MASS) has much potential to improve the efficiency of traditional apple breeding. Theoretical and empirical evaluation of selection cost and theoretical evaluation of genetic gain in previous studies support this promise. However, there is a lack of empirical validation of MASS efficiency reported in tree fruit crops for genetic gain and cost nor even genetic gain alone. To determine the optimal way of conducting MASS especially when genotypic information is less predictive than phenotypic information, which is the case for many quantitative traits, this study evaluated the genetic gain and cost of applying a multi-locus DNA test targeting fruit acidity, crispness, and firmness in a multi-family apple seedling population. Variance component analysis indicated that the DNA test explained 12%, 16%, and 17% of the additive variance of apple acidity, crispness, and firmness, respectively. Two-stage seedling selection was estimated to outperform traditional seedling selection for both genetic and cost efficiency: when 70% of the total seedlings were selected in the first stage based on marker score and 20% in the second stage based on phenotype, compared to phenotypic selection, two-stage selection identified 4% more individuals with breeding values within target ranges for acidity and firmness while achieving selection cost savings at DNA testing prices of less than $4.5/sample. This empirical validation of MASS suggested that when predictiveness of a DNA test is lower than heritabilities of the traits, properly weighting genotypic and phenotypic information through two-stage selection can lead to a higher genetic gain while reducing selection cost compared to phenotypic seedling selection.