Summary Flavor is a crucial aspect of the eating experience, reflecting evolving consumer preferences for fruits with enhanced quality. Modern fruit breeding programs prioritize improving quality traits aligned with consumer tastes. However, defining fruit‐quality attributes that significantly impact consumer preference is a current challenge faced by the industry and breeders. This study proposes a data‐driven approach to statistically model the relationship between fruit‐quality parameters and consumers' overall liking. Our primary hypothesis suggests that the interplay between fruit‐quality attributes and consumer preferences may reach a critical value, serving as new empirical benchmarks for fruit quality. Using extensive historical datasets accounting for sensory, biochemical, and genomic information described in blueberry, we first demonstrated that multivariate adaptive regression splines (MARS) could be used to identify specific values of fruit‐quality traits that significantly affect consumer perception by using nonlinear spline regressions on estimating threshold points. We harnessed genomic information and carried out genomic selection (GS) for five fruit‐quality traits evaluated on the original scale and after classified via the MARS approach. This study provides a pioneering consumer‐centric and data‐driven approach to defining fruit‐quality standards and supporting molecular breeding that has broad applications to breeding programs from any species.

中文翻译：

---

一种以消费者为导向的方法，用于定义分子育种的育种目标


总结 风味是饮食体验的一个重要方面，反映了消费者对质量更高的水果不断变化的偏好。现代水果育种计划优先考虑提高符合消费者口味的品质性状。然而，定义对消费者偏好有重大影响的水果质量属性是该行业和育种者当前面临的挑战。本研究提出了一种数据驱动的方法来对水果质量参数与消费者整体喜好之间的关系进行统计建模。我们的主要假设表明，水果品质属性与消费者偏好之间的相互作用可能达到临界值，作为水果品质的新实证基准。使用解释蓝莓中描述的感觉、生化和基因组信息的广泛历史数据集，我们首先证明了多变量自适应回归样条曲线 （MARS） 可用于识别显着影响消费者感知的水果品质性状的特定值通过使用非线性样条回归来估计阈值点。我们利用基因组信息，对在原始量表上评估并通过 MARS 方法分类后的五个水果品质性状进行了基因组选择 （GS）。这项研究提供了一种开创性的以消费者为中心和数据驱动的方法来定义水果质量标准并支持分子育种，该方法对任何物种的育种计划都有广泛的应用。