Abstract: Objective: To gain a deep understanding of the taste characteristics of bone peptides from different species of livestock and poultry and the differences between them, and to explore the feasibility of using electronic tongue in their rapid qualitative identification. Methods: Totally 16 samples of yak bone peptides (YBPs), bovine bone peptides (BBPs), pig bone peptides (PBPs) and chicken bone peptides (CBPs), four samples from each species, were collected for analysis of nutritional components and amino acid composition to clarify the material basis of the taste difference between them, and then the data of their taste characteristics were quantitatively discriminated by electronic tongue combined with multivariate statistical analysis. Results: The nutritional composition and amino acid composition of bone peptides from different animal species were different. Principal component analysis (PCA) and discriminant factor analysis (DFA) distinguished bone peptides from different animal species, and DFA was significantly more effective than PCA. Supervised partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA) and Fisher discriminant analysis model all had good discriminant capability for the taste characteristics of bone peptides, and the discriminant accuracy of the Fisher model was 68.8%. Conclusion: The taste characteristics of the four bone peptides differ umami, sweetness and bitterness. Electronic tongue taste detection combined with multivariate statistical analysis can provide a rapid and reliable means for the discrimination of bone peptides from different animal species.

Key words: electronic tongue; bone peptides from livestock and poultry; multivariate statistical analysis; discriminant analysis