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

Driven by the large amount of goat milk destined for cheese production, and to pioneer industry, objective this study was assess effect farm in predicting milk-coagulation curd-firmness traits via Fourier-transform infrared spectroscopy. Spectra from 452 Sarda goats belonging 14 farms central southeast Sardinia (Italy) were collected. A Bayesian linear regression model used, estimating all spectral wavelengths' effects simultaneously. Three traditional properties [rennet coagulation time (min), curd firmness 20 mm 30 min after rennet addition (mm)] 3 measures modeled over estimated according change (RCTeq), instant curd-firming rate constant, asymptotical firmness] considered. stratified cross validation (SCV) assigned, evaluating each separately (validation set; VAL) keeping remaining train (calibration set) statistical model. Moreover, a SCV, where 20% randomly taken (10 replicates per farm) VAL entered calibration set, also considered (SCV80). To performance, coefficient determination (R2VAL) root mean squared error recorded. The R2VAL varied between 0.14 0.45 (instant constant RCTeq, respectively), albeit standard deviation approximating half traits. Although average results 2 SCV procedures similar, SCV80, maximum increased at about 15% across traits, with highest observed (20%) lowest RCTeq (6%). Further investigation evidenced important variability among farms, some them being close 0. Our work outlined importance considering when developing spectroscopy prediction equations goats.