Mutton-like flavoring base was prepared from sheep bone meal by Maillard reaction. Sheep bone meal was sequentially hydrolyzed by alcalase and flavourzyme and was subjected to Maillard reaction with L-cysteine, L-glutamic acid, D-xylose and glucose. The effects of different reaction conditions (heating temperature, time and initial pH value) on the browning degree, sensory evaluation and color difference of Maillard reaction products were investigated. The reaction conditions were optimized by response surface methodology (RSM), and the volatile components of the flavoring base were analyzed by solid phase microextraction/gas chromatography-mass spectrometry (SPME/GC-MS). The results showed that the optimal reaction conditions were as follows: L-cysteine concentration 2 g/100 mL, L-glutamic acid concentration 3 g/100 mL, D-xylose concentration 2 g/100 mL, and glucose concentration 1 g/100 mL, reaction temperature 110 ℃, heating time 50 min, and initial pH 6.0. A total of 51 volatile substances were identified in the flavoring base, including 14 aldehydes, 7 alcohols, 6 ketones, 7 acids, 3 esters, one ether, 7 hydrocarbons, 3 sulfur-containing compounds and 3 heterocycles. The main substances contributing to the meaty aroma of the flavoring base were (E)-2-octenal, 1-octene-3-ol, (E)-2-nonenal and (E,E)-2,4-decadienal.

To develop an objective evaluation method for meat flavors, electronic nose responses were employed as objective evaluation measures to compensate for the disadvantages of traditional sensory evaluation, such as strong randomness as well as poor repeatability and stability. In this study, Maillard reaction was used to prepare two meat flavors: barbecue-like and broth-like, and the correlation between sensory evaluation and electronic nose responses was established by using grey correlation degree, principal component analysis (PCA), and analysis of variance (ANOVA). The results showed that through Maillard reaction with 10.00 g/100 mL of glutamic acid, a strong broth-like flavor was obtained, and the sensors W5S, W6S, W1S, W1W, and W2S had higher response values to the broth-like flavor, and their correlation coefficients with the sensory evaluation scores were 0.738, 0.803, 0.822, 0.825 and 0.864, respectively; when 3.33 g/100 mL of glutamic acid was added to the reaction system, a strong barbecue-like flavor was obtained, and the sensors W5S, W6S, W1S, and W1W had higher response values to the barbecue-like flavor, and their correlation coefficients with the sensory evaluation scores were 0.744, 0.835, 0.847, and 0.854, respectively. On this basis, a regression model was developed between the correlation coefficients of the sensor response values with higher grey correlation degree and the sensory scores, which had excellent goodness of fit.

In this study, the composition of volatile flavor substances during the fermentation of cream was analyzed by solid phase microextraction (SPME) and gas chromatography-mass spectrometry (GC-MS). A total of 32 volatile substances were identified at five different periods, among which acids were the most abundant. Principal component analysis (PCA) revealed that the composition of volatile compounds differed at different fermentation periods, and samples from different periods could be clearly distinguished. Nine key flavor components, including capric acid, caprylic acid, δ-decalactone, 2-nonanone and 3-hydroxy-2-butanone, were identified by calculating odor activity values (OAVs). Five sensory attributes of the samples were scored by descriptive sensory analysis. The correlations between key flavor substances and sensory attributes were analyzed by partial least squares regression (PLSR) analysis, and it was found that the fermented and creamy flavor had strong correlations with capric acid, capric acid, δ-decalactone, 2-decanone, 2-nonanone and 3-hydroxy-2-butanone. This study is of great significance to regulating the production of fermented cream.