Chen, Yangyi
(2020)
Modelling and predicting the quality of Cheddar cheese during ripening.
PhD thesis, University of Nottingham.
Abstract
Cheddar cheese maturation is crucial during cheese manufacture producing distinct flavours and textures. However, it is a slow and expensive process and also not all of the cheese production is capable of reaching the highest quality. Therefore, predictive tools are necessary to allow the optimisation and consistency of supply of mild-mature/vintage Cheddar cheeses, minimising losses.
Six blocks of Cheddar cheese produced on the same day in the same location were graded after a short period of storage by professional cheese graders and the Gilles and Lawrence grading model which is still used in cheese manufacture. After 13 months, these batches were regarded by graders again. Batch to batch variations of cheese ripening developments with different predictive quality were observed up to 450 days from state of water and fat, metabolic profile, aroma profile and texture profile. The batch to batch variation quality markers and ripening evolution markers from instrumental analysis were explored.
In parallel, sensory evaluation combined with the chemo-metrics was studied until 540 days ripening which is about the ripening time for the commercial vintage Cheddar cheese. Sensory profile identified that ‘downgrade’ batch C was found to have bland flavour and lumpy mouthfeel as well as yellower IV colour, and whose maturation level was behind schedule as determined sensorically. Whereas ‘downgrade’ batch E appeared to have strong flavour intensity, dirty aftertaste and texture defects. These suggested that Gilles and Lawrence model is not sufficient to predict the cheese quality.
Time domain water and fat proton Nuclear Magnetic Resonance signals are assigned. Transverse relaxation times (T2) for water and fat protons decrease and thermodynamic free water percentage increases with cheese ripening up to 450 days. Water and fat state attributes can differentiate between batches of Cheddar cheese after 56 days ripening.
Water soluble metabolites were extracted, identified and analysed using high resolution 1H13C Nuclear Magnetic Resonance experiments. The analytical methods coupled with chemo-metric analysis offer a profiling of metabolites which showed batch to batch variations and in addition showed a pathway among different ripening time points. Batch C revealed a higher level of serine and β-galactose as well as a lower amount of lactic acid in the aqueous extract. The normalised intensity of citrulline and arginine decreased during maturation.
The aroma profiles of the batches of cheese were studied by solid-phase microextraction gas chromatography-mass spectrometry during ripening. The trajectory of different predictive qualities of Cheddar cheese during ripening was presented. Secondary alcohols, propyl esters and acetic acid can be considered as a defect sign.
V Additionally, Protein matrix porosity analysed from microscopy images decreases during ripening with consequential decreases in cohesiveness and springiness.
The correlation between the instrumental analysis and sensory profile was studied. The normalised tyrosine, tyramine, lysine ratio in cheese aqueous extracts; acetoin and branch chain alcohols, and fracturability are highly correlated with a mature Cheddar cheese sensory profile. Conversely, glycerol, β-galactose content, springiness, protein matrix porosity and cohesiveness are associated with a young Cheddar cheese sensory profile. Texture related sensory attributes are correlated with octanoic acid, valeric acid and caproic acid levels. Cohesiveness is the attribute most correlated with sensory attributes among all texture profile analysis parameters.
Finally, a preliminary model was established and the sensory intensity of sweaty flavour, rate of melting, crumbly and onion flavour were well forecasted after 540 days ripening based on the 56 days measurements
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)