In-line near-infrared analysis of milk coupled with machine learning methods for the daily prediction of blood metabolic profile in dairy cattle

Abstract

Precision livestock farming technologies are used to monitor animal health and welfare parameters continuously and in real time in order to optimize nutrition and productivity and to detect health issues at an early stage. The possibility of predicting blood metabolites from milk samples obtained during routine milking by means of infrared spectroscopy has become increasingly attractive. We developed, for the first time, prediction equations for a set of blood metabolites using diverse machine learning methods and milk near-infrared spectra collected by the AfiLab instrument. Our dataset was obtained from 385 Holstein Friesian dairy cows. Stacking ensemble and multi-layer feedforward artificial neural network outperformed the other machine learning methods tested, with a reduction in the root mean square error of between 3 and 6% in most blood parameters. We obtained moderate correlations (r) between the observed and predicted phenotypes for γ-glutamyl transferase (r = 0.58), alkaline phosphatase (0.54), haptoglobin (0.66), globulins (0.61), total reactive oxygen metabolites (0.60) and thiol groups (0.57). The AfiLab instrument has strong potential but may not yet be ready to predict the metabolic stress of dairy cows in practice. Further research is needed to find out methods that allow an improvement in accuracy of prediction equations.

Figure 1

Pearson correlations heatmap between light emitting diodes (LEDs) information and blood metabolites in all Holstein–Friesian cows involved in the trial. On the left side, unsupervised hierarchical clustering of LEDs is showed as a dendrogram. NEFA non-esterified fatty acids, BHBA β-hydroxybutyric acid, AST aspartate aminotransferase, GGT γ-glutamyl transferase, BILt total bilirubin, ALP alkaline phosphatase, PON paraoxonase, ROMt total reactive oxygen metabolites, AOPP advanced oxidation protein products, FRAP ferric reducing antioxidant power, SHp thiolic groups, PROTt total proteins. Graphics have been created using the corrplot R package within the R software v. 3.6.3 (www.r-project.org).

Figure 2

Origin of substances present in mammary gland of dairy cows captured with near-infrared (NIR) spectra and relationships with blood stream and liver. The thick arrows represent gross transfer between compartments during lactation; in the mammary gland, the metabolites involved in these processes are mainly derived from blood flow. The thin arrows denote minor fluxes. Dashed lines represent fluxes of metabolites that are partly originated from the blood stream and partly secreted directly by mammary gland. During lactation, circulating non esterified fatty acids (NEFA) are regularly incorporated into milk fat; circulating NEFA in excess are stored in the liver as triglycerides (TG). Cholesterol (Chol) in the mammary gland originates from blood uptake and, to a certain extent, from local synthesis in the mammary tissue. Propionic acid represents the substrate for gluconeogenesis: the mammary gland absorbs large amounts of glucose from blood flow to synthetize lactose, enhance viability and proliferation of the mammary cells, and supply energy for synthesis of milk components. Acute phase proteins (APPs), including haptoglobin and ceruloplasmin, are mainly synthetized in the liver and are able to pass through the blood-mammary barrier in case of systemic inflammatory conditions or, when local processes are active, they could be self-produced by the mammary epithelium. ALP alkaline phosphatase, AST aspartate aminotransferase, BHBA β-hydroxybutyric acid, Crea creatinine, GGT γ-glutamyl transferase, HDL high density lipoproteins, MPO myeloperoxidase, PON paraoxonase, PROTt total proteins, ROMt total reactive oxygen metabolites, SHp thiolic groups, VLDL very low density lipoproteins.

Figure 3

The average value for AfiLab milk spectra (solid black line represents the average, the purple region represents values inside mean ± 3*SD, the blue line represents the maximum value, and orange line the minimum value) that belonged to the considered Holstein–Friesian cows’ population (n = 385). LED light emitting diode.