Abstract
Tomato (Solanum lycopersicum) is the most valuable fruit and horticultural crop species worldwide. Compared with the fruits of their progenitors, those of modern tomato cultivars are, however, often described as having unsatisfactory taste or lacking flavor. The flavor of a tomato fruit arises from a complex mix of tastes and volatile metabolites, including sugars, acids, amino acids, and various volatiles. However, considerable differences in fruit flavor occur among tomato varieties, resulting in mixed consumer experiences. While tomato breeding has traditionally been driven by the desire for continual increases in yield and the introduction of traits that provide a long shelf-life, consumers are prepared to pay a reasonable premium for taste. Therefore, it is necessary to characterize preferences of tomato flavor and to define its underlying genetic basis. Here, we review recent conceptual and technological advances that have rendered this more feasible, including multi-omics-based QTL and association analyses, along with the use of trained testing panels, and machine learning approaches. This review proposes how the comprehensive datasets compiled to date could allow a precise rational design of tomato germplasm resources with improved organoleptic quality for the future.