Merging fast sensory profiling with non-targeted GC-MS analysis for multifactorial experimental wine making
Abstract
Wine aroma is influenced by several viticultural and oenological factors. In this study we used experimental wine making in a full factorial design to determine the impact of grapevine age, must turbidity, and yeast strain on the aroma of Vitis vinifera L. cv. Riesling wines. A recently developed, non-targeted SPME-GC-MS fingerprinting approach for wine volatiles was used. This approach includes the segmentation and mathematical transformation of chromatograms in combination with Parallel Factor Analysis (PARAFAC) and subsequent deconvolution of important chromatogram segments. Additionally, fast sensory screenings of the experimental wines were conducted using a napping approach with free choice descriptor profiling. Experimental wine making using a full factorial design allowed for the determination of the main and interaction effects of the examined factors on the aroma and volatile composition of the wines. Groupings of the wines obtained from perceptual mapping could be correlated to volatile compounds and sensory descriptor groups using Multiple Factor Analysis (MFA), a multi-block PCA technique with a special scaling of the data blocks. The results provide a new analytical insight on the impact of the factors grapevine age, must turbidity and yeast strain on wine aroma. The analytical methodology and the data analysis approach presented here promise to shed new light on viticultural and oenological factors influencing wine aroma during the wine making process.
Issue: Macrowine 2016
Type: Poster
Authors
*Université de Bordeaux