2025 Volume 96 Issue 1 Pages 7-18
Japanese Black cattle have successfully differentiated themselves from imported beef through improved marbling. Consequently, grading has become highly standardized, necessitating new objective indicators. Although the composition of fatty acids is increasingly used as an indicator, the relationship between other components and eating quality remains unclear. Analytical sensory evaluation involves directly and objectively assessing eating quality; however, it is not suitable in case of multiple samples. In this study, a method of predicting eating quality characteristics from component analysis values that can be applicable in case of multiple samples was developed. Twelve attributes of sensory evaluation were considered, including “tenderness (bite)” (i.e., tenderness when the sample was bitten off for the first time), “tenderness (chew)” (i.e., tenderness when the sample was chewed in the mouth), “weakness of fibrousness”, “juiciness”, “greasiness”, “sweet flavor”, “grilled/roasted flavor”, “weakness of off-flavor”, “flavor intensity”, “umami intensity(IMP)” (i.e., taste like 5’-inosine monophosphate ), “umami intensity(Glu)” (i.e., taste like monosodium glutamate) and “umami intensity(total)”. Principal component analysis was performed to examine the sensory evaluation scores. In the results, the first to third principal components explained >93% of all information. Each principal component score was interpreted as “overall evaluation focusing on texture,” “taste” and “smell.” By using genetic algorithms and partial least squares regression analysis, formulas were created to predict sensory evaluations from component analysis values, with determination coefficients of 0.6-0.8 for many attributes. These results enable data collection on the taste quality of a sufficient number of samples to evaluate brand characteristics and breeding values, contributing to the understanding and improvement of taste quality in various brands.