Multivariate modelling of relationships between proteolysis and sensory quality of Prast cheese

Sammendrag

A study was undertaken to evaluate how the proteolytic processes during cheese ripening (i.e., cleavage of native caseins, formation of medium-sized and smaller peptides and free amino acids) have an impact on the development of mouth feel/texture and taste/aroma of maturing cheeses. Multivariate modelling of relationships between results from biochemical analyses of proteolysis and from sensory profiling of Prast samples (N = 15) were done using cross-validated partial least square regressions. Differences in salt perception mainly due to different salt content were correlated with several proteolytic data. Models based on profiles of smaller peptides found by reverse-phase chromatography of ethanol (70%)-soluble fractions had a correlation with descriptors such as flavour intensity (R = 0.83), off-flavour (R = 0.86) and bitterness (R = 0.72). Composition of free amino acid correlated with flavour intensity (R = 0.82), but significant correlations for more taste-related descriptors such as off-flavour and bitterness were not found. Multivariate modelling was a useful approach to relate biochemistry of cheese ripening to cheese quality as assessed by sensory profiling. (c) 2005 Elsevier Ltd. All rights reserved. A study was undertaken to evaluate how the proteolytic processes during cheese ripening (i.e., cleavage of native caseins, formation of medium-sized and smaller peptides and free amino acids) have an impact on the development of mouthfeel/texture and taste/aroma of maturing cheeses. Multivariate modelling of relationships between results from biochemical analyses of proteolysis and from sensory profiling of Präst samples (N=15) were done using cross-validated partial least square regressions. Differences in salt perception mainly due to different salt content were correlated with several proteolytic data. Models based on profiles of smaller peptides found by reverse-phase chromatography of ethanol (70%)-soluble fractions had a correlation with descriptors such as flavour intensity (R=0.83), off-flavour (R=0.86) and bitterness (R=0.72). Composition of free amino acid correlated with flavour intensity (R=0.82), but significant correlations for more taste-related descriptors such as off-flavour and bitterness were not found. Multivariate modelling was a useful approach to relate biochemistry of cheese ripening to cheese quality as assessed by sensory profiling.