The good, the bad and the aged: predicting sensory quality of anhydrous milk fat by PTR/SRI-Tof-MS analysis and data mining

Due to its versatility, anhydrous milk fat (AMF) has become more popular as a food industry ingredient,but its quality control remains a critical challenge. A direct injection mass spectrometry technique wasapplied to predict sensory quality of AMF. Volatilome analysis through proton transfer reaction massspectrometry (PTR-MS) was used to classify 39 industrial samples of AMF according to industrial sensoryevaluation and to accelerated ageing. A selective reagent ion system was used to evaluate the suitabilityof PTR-MS alternative ionisation modes for quality control. Supervised multivariate data analysis suc-cessfully classified samples and showed that samples exposed to accelerated shelf life at 50C presentedhigher intensities of most volatiles, especially for the ones derived from oxidation like aldehydes andketones, while samples with an acceptable quality level had lower emissions of volatiles. PTR-MStechnique is ideal to support agroindustry sensory quality programs requiring rapid on-line analytical informatio