Investigation of Components that Affect Flavors and Visualizing Differences in Tastes

Significance of the Topic

Understanding the molecular basis of complex sensory attributes in food and beverages is essential for objective quality control and product development. In sake production, the elusive sensation known as fukurami—where flavor expands in the mouth—cannot be fully captured by standard tastes alone. Integrating instrumental analysis with sensory evaluation enables identification of specific chemical markers underlying such ambiguous characteristics.

Study Objectives and Overview

This investigation aimed to establish a reproducible workflow for correlating sensory descriptors with instrumental data, focusing on fukurami in sake. Eight sake samples, adjusted to a uniform alcohol content (15 %), were brewed using different polishing rates and yeast strains. A trained panel ranked samples for fukurami, grouping four as “with” and four as “without.” Two borderline samples were set aside for model validation after initial analysis.

Methodology and Workflow

Preparations included 10× dilution of sake for flavor component analysis by ion-pair-free LC/MS/MS (targeting 153 hydrophilic metabolites) and headspace GC/MS for 21 volatile aroma compounds. Data preprocessing employed area ratios against internal standards. Multivariate analyses—principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA)—were performed to visualize patterns and extract discriminant markers. A support vector machine-based model was built using selected compounds to classify unknown samples.

Instrumentation Used

• Shimadzu LCMS-8060 triple quadrupole MS with Nexera HPLC
• Shimadzu GCMS-QP2020 with HS-20 headspace sampler
• Discovery HS F5-3 HPLC column (2.1×150 mm, 3 µm)

Main Findings and Discussion

PCA revealed clustering by yeast strain but did not align with fukurami perception. PLS-DA identified several flavor markers distinguishing fukurami groups. Key differentiators for “with fukurami” samples included higher relative levels of disaccharides (sweetness) and aroma compounds such as isobutanol, isobutyl acetate, and ethyl acetate. Conversely, “without fukurami” samples showed elevated organic acids (e.g., malic acid), certain amino acids, and nucleobases. Statistical tests (t-test, Mann-Whitney U) confirmed the significance of these components (p < 0.05). Discriminant models accurately classified the two validation samples according to their sensory grouping.

Benefits and Practical Applications

• The workflow offers objective markers for fukurami, guiding sake brewing and rice polishing decisions.
• Quantitative monitoring of flavor balance supports consistent quality control.
• Approach can be extended to other foods and beverages to characterize complex sensory traits beyond basic tastes.

Future Trends and Applications

• Integration of additional omics layers (metabolomics, proteomics) to refine flavor indices.
• Advanced machine learning for predictive sensory modeling and real-time quality monitoring.
• Application of the workflow to dynamic flavor perception studies and personalized taste profiling.

Conclusion

This study establishes a robust analytical framework linking sensory descriptors of fukurami with specific chemical constituents. Identification of sweetness- and aroma-related markers, coupled with discriminant modeling, enables reliable prediction and control of this complex mouthfeel attribute, enhancing sake quality management and broadening applications in food sensory science.

Reference

• Ozaki K, et al. Journal of the Brewing Society of Japan. 2008;103(3):150–162.
• Furukawa H, et al. Journal of the Brewing Society of Japan. 1983;78(6):419–422.
• Yoshizawa K, Koizumi T. Journal of the Brewing Society of Japan. 1997;92(3):217–223.
• Yoshizawa K. Journal of the Brewing Society of Japan. 1980;75(6):451–457.