Prediction of sensory characteristics of cider according to their biochemical composition: Use of a central composite design and external validation by cider professionals

- Predictive models of four attributes were built depending on cider composition.
- The experimental design enables a large overview of the main cider components.
- External validation was performed by professionals.

A large experimental design containing 72 cider-like model solutions was developed in order to evaluate the impact of apple procyanidins, fructose, acid content and ethanol on four sensory attributes: bitterness, astringency, sweetness and sourness. Based on cider composition, this work investigated the linear and quadratic effects of these four quantitative factors and the interactions between them. The results underline the role of procyanidin content (procyanidin profile and concentration) in cider, which highly impacts bitterness and astringency but also affects sweetness and sourness. Using an external validation based on a sensory evaluation by cider professionals, one predictive model for each sensory attribute is proposed.