Goat farm variability affects milk Fourier-transform infrared spectra used for predicting coagulation properties

Abstract

Driven by the large amount of goat milk destined for cheese production, and to pioneer the goat cheese industry, the objective of this study was to assess the effect of farm in predicting goat milk-coagulation and curd-firmness traits via Fourier-transform infrared spectroscopy. Spectra from 452 Sarda goats belonging to 14 farms in central and southeast Sardinia (Italy) were collected. A Bayesian linear regression model was used, estimating all spectral wavelengths' effects simultaneously. Three traditional milk-coagulation properties [rennet coagulation time (min), time to curd firmness of 20 mm (min), and curd firmness 30 min after rennet addition (mm)] and 3 curd-firmness measures modeled over time [rennet coagulation time estimated according to curd firmness change over time (RCT eq ), instant curd-firming rate constant, and asymptotical curd firmness] were considered. A stratified cross validation (SCV) was assigned, evaluating each farm separately (validation set; VAL) and keeping the remaining farms to train (calibration set) the statistical model. Moreover, a SCV, where 20% of the goats randomly taken (10 replicates per farm) from the VAL farm entered the calibration set, was also considered (SCV 80 ). To assess model performance, coefficient of determination (R 2 VAL ) and the root mean squared error of validation were recorded. The R 2 VAL varied between 0.14 and 0.45 (instant curd-firming rate constant and RCT eq , respectively), albeit the standard deviation was approximating half of the mean for all the traits. Although average results of the 2 SCV procedures were similar, in SCV 80 , the maximum R 2 VAL increased at about 15% across traits, with the highest observed for time to curd firmness of 20 mm (20%) and the lowest for RCT eq (6%). Further investigation evidenced important variability among farms, with R 2 VAL for some of them being close to 0. Our work outlined the importance of considering the effect of farm when developing Fouriertransform infrared spectroscopy prediction equations for coagulation and curd-firmness traits in goats.

Low Effectiveness of Mid-Infrared Spectroscopy Prediction Models of Mediterranean Italian Buffalo Bulk Milk Coagulation Traits

“…The 3.38% of the samples did not coagulate within the analysis. Other authors have also reported a greater presence of noncoagulating samples when dealing with individual milk from buffalo (16.9%) [11], sheep (12.9%) [9], and goats (7.9%) [8]. Regarding buffalo milk coagulation properties (Table 1), a previous study reported a shorter RCT (13.14 min) but a longer k 20 (4.89 min) with a thicker a 30 (48.32 mm) in bulk milk [4].…”

“…Moreover, these results are consistent with the use of MIR spectroscopy to predict milk coagulability across different types of milk. For instance, goat milk MIR prediction models developed with individual milk samples also reported a low R 2 ExV for all three traits [8]. While the R 2…”

“…Studies conducted with the individual milk of goats [8] and sheep [9] and with bulk milk from cow herds [10] have revealed the low accuracy of the prediction models for milk coagulation properties. To our knowledge, only one study conducted in 2017 has focused on its applicability to predict the coagulation properties of individual milk samples [11].…”

A Comprehensive Analysis and Detection Methodology Using Near-Infrared (NIR) Spectroscopy to Unveil the Deceptive Practice of Milk Adulteration

“…[10]. The process of cheese-making is complex, which has been well described by authors Dadausis et al, occurring through a coagulation technique called milk coagulation properties [11].…”

Use of Milk Infrared Spectral Data as Environmental Covariates in Genomic Prediction Models for Production Traits in Canadian Holstein

“…Wang and Bovenhuis [37] reported that across-herd predictions of bovine methane emissions showed much lower accuracy than within-herd predictions. Dadousis et al [40], in predicting goat milk coagulation traits, showed that model predictive ability depended largely on the farm(s) included in the validation set.…”