Mid-infrared spectrometry prediction of the cheese-making properties of raw Montbeliarde milks from herds and cheese dairy vats used for the production of Protected Designation of Origin and Protected Geographical Indication cheeses in Franche-Comte

摘要

Franche-Comte is the primary producing region ofProtected Designation of Origin cheeses in France.Normally, mid-infrared (MIR) prediction models forcheese-making property (CMP) traits are developedusing individual bovine milks. However, considering therequests of all actors in the dairy sector, the presentstudy aimed to assess the feasibility of MIR spectroscopyto develop CMP equations of Montbéliarde herd anddairy vat milks. For this purpose, 22 CMP traits wereanalyzed on samples collected in 2016 (half in February–March and half in May–June) from 100 commercialherds and 70 dairy vats (55 cheese dairies) located inFranche-Comté. These characteristics included 11 rennetcoagulation traits and 8 lactic acidification traitsmeasured in either soft cheese or pressed cooked cheeseconditions and 3 laboratory curd yields. Models of MIRprediction for each of the 22 CMP traits were built usingpartial least squares regression with external validationby dividing the data set into calibration (70%) andvalidation (30%) sets. We confirmed that the variabilityof milk traits depends largely on the production scaleand is higher for individual milk than for herd milk andeven higher for vat milk. The best prediction modelswere obtained in herd milk samples for curd yieldsexpressed in dry matter or fresh, with a coefficient ofdetermination (R~2) in external validation of 0.78 and0.77, respectively. As with individual milk, these traitsare closely related to the gross composition of the milkand therefore easier to predict by MIR spectroscopy.However, these curd yield traits were poorly predicted(R~2 = 0.58) in vat milk samples due to their lowervariability. In herd milk samples, prediction models ofother CMP traits were poorly accurate except for theratio of the time to obtain a standard firmness to therennet coagulation time in soft cheese or pressed cookedcheese conditions, which showed R~2 ＞ 0.66 in externalvalidation. Such trait is important in qualifying thebehavior of milk during cheese production. Predictionmodels of other CMP traits for either herd or vat milksamples had poor accuracy, and further work is neededto improve their performance.