Prediction of sweet pepper (Capsicum annuum) flavor over different harvests

To better understand and predict the complex multifactorial trait flavor, volatile and non-volatile components were measured in fresh sweet pepper (Capsicum annuum) fruits throughout the growing season in a diverse panel of 24 breeding lines, hybrids, several cultivated genotypes and one gene bank accession. Biochemical profiles were linked to individual flavor attributes, that were objectively quantified by a trained descriptive expert panel. We used a Random Forest regression approach for prediction of the flavor attributes within and between harvests. Predictions of texture related attributes (juiciness, toughness, crunchiness and stickiness of the skin) and sweetness were good (around 60-65 % in the analyses with the three harvests combined). The predictions of the attributes aroma intensity, sourness and fruity/apple were somewhat lower and more variable between harvests. (E)-2-hexen-1-ol, neopentane, p-menth-1-en-9-al, 3-hepten-2-one, (Z)-beta-ocimene, (Z)-2-penten-1-ol, 1-methyl-1,4-cyclohexadiene, glucose, fructose and three unknown volatile compounds were identified as key-metabolites involved in the flavor differences between both genotypes and harvests. The complex nature of flavor is exemplified by the observed masking effect of fructose and other sugars on sourness and sourness related metabolites, like citrate. The knowledge obtained from the overall biochemical, sensory and prediction analyses forms a basis for targeted flavor improvement by breeding.