HS-SPME-GC-MS combined with multivariate statistical analysis reveals off-flavor composition and biomarker generation mechanism of whole milk powder

Milk powder is prone to producing off-flavor when it is not handled properly during production or storage. In order to identify the off-flavor compounds of whole milk powder (WMP), headspace solid-phase microextraction gas chromatography-mass spectrometry (HS-SPME-GC-MS) combined with cluster heatmap analysis was used, orthogonal partial least squares discriminant analysis (OPLS-DA) and principal component analysis (PCA) were used to investigate the off-flavor substances of whole milk powder. Cluster heatmap analysis showed that there were certain differences in the relative contents of volatiles among the eight milk powders. Further, OPLS-DA was applied to the relative contents of volatile substances in milk powders with and without unpleasant odor, a to obtain the main difference components of the 28 with variable importance in projection (VIP) greater than 1, with off-flavor substances being the main ones. Finally, the volatile compounds positively correlated with off-flavors in milk powder were screened by principal component analysis and combined with relative aroma activity values greater than 1 to identify (E)-2-hexenal and (E,E)-2,4-decadienal as the major off-flavor compounds of milk powder, and also described the mechanism of generation of (E)-2-hexenal (a marker of soya flavor) and (E,E)-2,4-decadienal (a marker of grass flavor) in milk powder. The results provide a reference for the analysis of flavor profile and possible off-flavor components of milk powder, which is beneficial for the optimization and improvement of whole milk powder.