Genetic and environmental variation in bovine milk infrared spectra

Abstract

Fourier transform infrared (FTIR) spectroscopy is widely used to determine milk composition. In this study, 1,060 milk infrared wavenumbers ranging from 925 to 5,008cm(-1) of 1,748 Holstein Friesian cows on 371 herds in the Netherlands were available. The extent to which infrared wavenumbers are affected by genetic and environmental factors was investigated. Inter-herd heritabilities of 1,060 infrared wavenumbers ranged from 0 to 0.63, indicating that the genetic background of infrared wavenumbers differs considerably. The majority of the wavenumbers have moderate to high inter-herd heritabilities ranging from 0.20 to 0.60. The diacylglycerol O-acyltransferase 1 (DGAT1), stearoyl-CoA desaturase (SCD1), κ-casein (CSN3), and β-lactoglobulin (LGB) polymorphisms are known to have a large effect on milk composition, and therefore we studied the effects of these polymorphisms on infrared wavenumbers. The DGAT1 polymorphism had highly significant effects on many wavenumbers. In contrast, the SCD1 polymorphism did not significantly affect any of the wavenumbers. The SCD1 is known to have a strong effect on the content of C10:1, C12:1, C14:1, and C16:1 fatty acids. Therefore, these results suggest that FTIR spectra contain little direct information on these monounsaturated fatty acids. The CSN3 and LGB polymorphisms had significant effects on a few wavenumbers. Differences between herds explained 10 to 25% of the total variance for most wavenumbers. This suggests that the wavenumbers of milk FTIR spectra are indicative for differences in feeding and management between herds. The wavenumbers between 1,619 and 1,674cm(-1) and between 3,073 and 3,667cm(-1) are strongly influenced by water absorption and usually excluded when setting up prediction equations. However, we found that some of the wavenumbers in the water absorption region are affected by the DGAT1 polymorphism and lactation stage. This suggests that these wavenumbers contain useful information regarding milk composition.

A landscape of the heritability of Fourier-transform infrared spectral wavelengths of milk samples by parity and lactation stage in Holstein cows

“…Group 1 included the SWIR-MWIR and the MWIR2 regions; these regions (which are associated with absorbance by water molecules) have very high phenotypic variation (Supplemental Figure S2; https: / / doi .org/ 10 .3168/ jds .2018 -15109) and very low proportion of variance explained both by genetic factors and by herd management practices and seasonal differences (Figure 1). Our heritability estimates for these 2 regions agree with results reported previously (e.g., Soyeurt et al, 2010;Bittante and Cecchinato, 2013;Wang et al, 2016). However, toward the right-hand limit of MWIR2 region, near wavenumber 1,550 cm −1 , there was a clear estimated increase in heritabilities.…”

“…Our heritability estimates in the SWIR region were similar to those reported by Bittante and Cecchinato (2013) for Brown Swiss and by Soyeurt et al (2010) for Belgian Holstein. Using data from Dutch Holsteins, Wang et al (2016) reported much higher heritability estimates for this region. The results reported by Wang et al (2016) for this region of the spectra are unusually high compared with those of previous studies and our own results.…”

“…Here, I is an identity matrix, A is the pedigree relationship matrix, σ hys 2 is the variance of the herd-year-season effects, σ a 2 is the additive genetic variance, and σ ε 2 is the error variance. To obtain results comparable with those who have previously studied the heritability of FTIR milk spectral data (e.g., Soyeurt et al, 2010;Bittante and Cecchinato, 2013;Wang et al, 2016;Wang and Bovenhuis, 2018), we defined heritability as the inter-herd proportion of variance explained by additive genetic effects; that is, Variance components and heritabilities were estimated for each trait (either absorbance at one of the 1,060 wavelengths, or MY, Fat%, or Prot%) for each stratum separately, resulting in a total of 19,134 different analyses.…”

“…Developing an optimal strategy for the use of milk FTIR data in genetic evaluations requires knowledge 1355 of the heritability of absorbance at individual wavelengths. A few authors have estimated genetic parameters of individual milk FTIR wavelength absorbance data (e.g., Soyeurt et al, 2010;Bittante and Cecchinato, 2013;Wang et al, 2016;Wang and Bovenhuis, 2018). These studies have shown that absorbance (and transmittance) at certain regions of the spectrum can be moderately heritable.…”

Standardization of milk mid-infrared spectrometers for the transfer and use of multiple models

“…A set of standardization samples were measured on a reference instrument (the "master" instrument) and on each instrument that needs to be aligned (the "slave" instruments), leading to different response matrices. As reported by Grelet et al (2015) and Wang et al (2016), individual FT-MIR spectrometers suffer from instability over time. To cope with this instability, the master cannot be a single instrument.…”