Parent Selection in Breeding for Roasted Peanut Flavor Quality1

Abstract

The sensory attributes that make up roasted peanut flavor quality are important traits to evaluate in the development of new cultivars. Recent publications have characterized the variation in sensory attributes in U.S. peanuts (Arachis hypogaea L.), however, no estimates of the effects of lines as parents in a breeding program have been calculated. Best linear unbiased prediction (BLUP) is a method for predicting the breeding value of a parent based on the performance of its relatives. Commonly used in animal and tree breeding, the method is rarely applied in annual crop species. The method was applied to a set of data collected on the three sensory attributes roasted peanut, sweet, and bitter for 250 peanut genotypes evaluated in 53 environments. BLUP solutions computed using different estimates of narrow-sense heri-tability (h2) were highly correlated (r > 0.9), suggesting that precise estimates of h2 are not necessary. Correlations of values predicted by BLUP with observed values were moderate (0.63 < r < 0.71) for individual lines, but strong (0.85 < r < 0.92) for means of crosses. BLUPs of breeding value can be used to predict means of crosses, but segregation within crosses affords additional opportunity for progress from selection. BLUPs of breeding value were superior to midparent values in predicting cross means. The previously reported flavor relationships between the runner and virginia market types were predicted by the breeding values of some commonly used parents. Specific lines with superior breeding values for flavor quality included runner-types Florunner, its components and sibling lines, and high-oleic derivatives of Sunrunner; virginia-types Altika, White's Runner, NC Ac 18457, and X90053; and spanish-type Pearl. Jenkins Jumbo, Florigiant, NC7, and CBR-resistant lines had inferior breeding values for flavor quality.