Importance of Mid-Infrared Spectra Regions for the Prediction of Mastitis and Ketosis in Dairy Cows

SIMPLE SUMMARY: Mid-infrared spectral regions are routinely used to predict milk components, such as fat and protein, but can also be assisting to detect diseases in dairy cows, including mastitis and ketosis. However, not all wavelengths carry relevant information and are therefore not necessary for prediction. Thus, the objective of this study was to identify which wavelengths are of particular importance for prediction of mastitis and ketosis. Results indicate that important wavenumbers varied across different traits. For the prediction of mastitis, 23 wavelengths were selected as highly relevant, while for the prediction of ketosis, 61 wavelengths were particularly important. Thus, this implies that it is reasonable to select the most important spectral regions for the prediction of mastitis and ketosis in dairy cows. ABSTRACT: Mid-infrared (MIR) spectroscopy is routinely applied to determine major milk components, such as fat and protein. Moreover, it is used to predict fine milk composition and various traits pertinent to animal health. MIR spectra indicate an absorbance value of infrared light at 1060 specific wavenumbers from 926 to 5010 cm(−1). According to research, certain parts of the spectrum do not contain sufficient information on traits of dairy cows. Hence, the objective of the present study was to identify specific regions of the MIR spectra of particular importance for the prediction of mastitis and ketosis, performing variable selection analysis. Partial least squares discriminant analysis (PLS-DA) along with three other statistical methods, support vector machine (SVM), least absolute shrinkage and selection operator (LASSO), and random forest (RF), were compared. Data originated from the Austrian milk recording and associated health monitoring system (GMON). Test-day data and corresponding MIR spectra were linked to respective clinical mastitis and ketosis diagnoses. Certain wavenumbers were identified as particularly relevant for the prediction models of clinical mastitis (23) and ketosis (61). Wavenumbers varied across four distinct statistical methods as well as concerning different traits. The results indicate that variable selection analysis could potentially be beneficial in the process of modeling.