Metabolomic modelling of sensory characteristics and consumer liking in papaya fruit

Specific flavour traits in papaya significantly influence consumer acceptance, with unfavourable flavours limiting marketability. However, limited evidence links metabolites to sensory characteristics in papaya. This study characterised flavour profiles of 27 papaya genotypes from Australia and South-east Asia (nine commercial cultivars, 18 breeding lines), using metabolomic and sensory data. Ten trained panellists performed descriptive analysis using 20-attribute descriptors, and 125 consumers evaluated flavour acceptance. Variance decomposition analysis revealed that aroma characteristics and volatile organic compounds accounted for most of the variance in consumer liking (31% and 63%, respectively). Ensemble machine learning determined that lipid-derived compounds (e.g. γ-octalactone) and terpenoids (e.g. linalool oxide) influenced the sensory attributes and fruit liking. Gradient boosting regression predicted consumer acceptance using three key metabolites: sulcatone (81%, variable importance), Brix (10%), and β-ionone (9%). These findings provide a novel approach to quantify fruit flavour in papaya, facilitating the development of new cultivars.