From the current probabilistic approach to a deterministic production process, a clear step towards digital transformation in the wine sector

Abstract

Currently, to consistently ensure the maintenance of a wine-style while benefiting from the utmost rigor made possible by the winemaking process, the composition of the wine blend is made using sensory control. This is performed after the wine is made with no real possibility of deterministic intervention.

However, different sensory profiles are frequently observed in high-volume wines fermented from the same grape must batch under similar fermentation conditions. So far, it has not been possible, using winery-available resources, to understand the drivers of these differences. Moreover, the impact of using a sulphitation / desulphitation process of stabilization on the varietal aromatic potential of white must is unknown.

Indeed, for modeling sensory evolution from must to wine, it is necessary to know, quite in-depth, the chemical a specific metabolic state characteristic of each style (this being a sensory profile) and study the relationships between sensory descriptors and volatile compounds described to be markers of sensory typicality (key-compounds) in determining high-volume mass-market standard profiles.

In this study, the varietal aromatic potential of 1 ML of white grape must was characterized during sulphitation / desulphitation process, together with the resulting wines, fermented by a controlled process using sensors to measure different key-parameters. The volatile composition and glycosylated fraction were studied by comprehensive two-dimensional chromatography (GC x GC-ToFMS) according to procedures previously implemented in the x-Chromatography Lab (http://xchromatographylab.x10.mx/). Obtained results indicate that the stabilization process by sulphitation / desulphitation has a highly significant impact on must composition, resulting from a complex network of effects. Five main groups of aroma descriptors were used in the sensory analysis of wines resulting from this project (herbaceous / vegetable, citrus, tropical fruits, orchard fruits and floral). Two-part information networks were built combining chemical and sensory information. It was possible to observe a grouping of experimental wines for the same binomen must origin / yeast. In addition, each wine displayed individuality with respect to sensory analysis, volatile profile and physicochemical parameters.

This work involves a new concept, still underexploited in the wine sector, promising to totally revolutionize classical concepts of oenology by integrating, in the same pro