Testing and developing the tomato flavour model

Regularly testing the tomato flavour model

The flavour model of Wageningen University & Research is a reliable tool to determine flavour levels of new tomato varieties. It facilitates B2B communication on flavour and thus contributes to transparency and trust in the chain. The algorithm of the model that calculates the flavour from the underlying input parameters has evolved over the years through calibration experiments. Regular calibration tests are necessary to ensure that current developments in breeding and the development of flavour perception of consumers are properly reflected in the model prediction. In September 2018 a new calibration experiment was held. Users of the model sent in tomato samples of different flavour levels. A large array of different flavour levels was presented to the panel. The results showed that the present model is still a good and reliable predictor of flavour levels. In an experiment with 18 varieties of different flavour levels, the model predicted 92% of the flavour differences. We concluded that adaptation of the model is not necessary. In the spring of 2019, the calibration experiment will be repeated.

Developing a successor

In February 2019 also a new project will start. Based on our existing knowledge and experience we will develop a new flavour model for tomato with non-destructive input parameters. This method is to use a handheld and non-destructive sensor, predicting the outcome of a hedonic panel from direct measurements. This can be applied in every greenhouse, every warehouse and every shop environment. Companies do not have to wait until the new technique has been fully developed: They can still use the current model predictions for flavour levels to guarantee and develop the flavour of fresh tomatoes.