Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations.

At the individual cow level, suboptimum fertility, mastitis, negative energy balance, and ketosis are major issues in dairy farming. These problems are widespread on dairy farms and have an important economic impact. The objectives of this study were (1) to assess the potential of milk mid-infrared (MIR) spectra to predict key biomarkers of energy deficit (citrate, isocitrate, glucose-6 phosphate [glucose-6P], free glucose), ketosis (ß-hydroxybutyrate [BHB] and acetone), mastitis (N-acetyl-ß-d-glucosaminidase activity [NAGase] and lactate dehydrogenase), and fertility (progesterone); (2) to test alternative methodologies to partial least squares (PLS) regression to better account for the specific asymmetric distribution of the biomarkers; and (3) to create robust models by merging large datasets from 5 international or national projects. Benefiting from this international collaboration, the dataset comprised a total of 9,143 milk samples from 3,758 cows located in 589 herds across 10 countries and represented 7 breeds. The samples were analyzed by reference chemistry for biomarker contents, whereas the MIR analyses were performed on 30 instruments from different models and brands, with spectra harmonized into a common format. Four quantitative methodologies were evaluated to address the strongly skewed distribution of some biomarkers. Partial least squares regression was used as the reference basis, and compared with a random modification of distribution associated with PLS (random-downsampling-PLS), an optimized modification of distribution associated with PLS (KennardStone-downsampling-PLS), and support vector machine (SVM). When the ability of MIR to predict biomarkers was too low for quantification, different qualitative methodologies were tested to discriminate low versus high values of biomarkers. For each biomarker, 20% of the herds were randomly removed within all countries to be used as the validation dataset. The remaining 80% of herds were used as the calibration dataset. In calibration, the 3 alternative methodologies outperform the PLS performances for the majority of biomarkers. However, in the external herd validation, PLS provided the best results for isocitrate, glucose-6P, free glucose, and lactate dehydrogenase (coefficient of determination in external herd validation [R2v] = 0.48, 0.58, 0.28, and 0.24, respectively). For other molecules, PLS-random-downsampling and PLS-KennardStone-downsampling outperformed PLS in the majority of cases, but the best results were provided by SVM for citrate, BHB, acetone, NAGase, and progesterone (R2v = 0.94, 0.58, 0.76, 0.68, and 0.15, respectively). Hence, PLS and SVM based on the entire dataset provided the best results for normal and skewed distributions, respectively. Complementary to the quantitative methods, the qualitative discriminant models enabled the discrimination of high and low values for BHB, acetone, and NAGase with a global accuracy around 90%, and glucose-6P with an accuracy of 83%. In conclusion, MIR spectra of milk can enable quantitative screening of citrate as a biomarker of energy deficit and discrimination of low and high values of BHB, acetone, and NAGase, as biomarkers of ketosis and mastitis. Finally, progesterone could not be predicted with sufficient accuracy from milk MIR spectra to be further considered. Consequently, MIR spectrometry can bring valuable information regarding the occurrence of energy deficit, ketosis, and mastitis in dairy cows, which in turn have major influences on their fertility and survival.

Imputation of missing milk Fourier transform mid-infrared spectra using existing milk spectral databases: A strategy to improve the reliability of breeding values and predictive models.

The use of milk Fourier transform mid-infrared (FT-MIR) spectrometry to develop management and breeding tools for dairy farmers and industry is growing and supported by the availability of numerous new predicted phenotypes to assess the nutritional quality of milk and its technological properties, but also the animal health and welfare status and its environmental fingerprint. For genetic evaluations, having a long-term and representative spectral dairy herd improvement (DHI) database improves the reliabilities of estimated breeding values (EBV) from these phenotypes. Unfortunately, most of the time, the raw spectral data used to generate these estimations are not stored. Moreover, many reference measurements of those phenotypes, needed during the FT-MIR calibration step, are available from past research activities but lack spectra records. So, it is impossible to use them to improve the FT-MIR models. Consequently, there is a strong interest in imputing those missing spectra. The innovative objective of this study was to use the existing large spectral DHI database to estimate missing spectra by selecting probable spectra using, as the match criteria, common dairy traits recorded for a long time by DHI organizations. We tested 4 match criteria combinations. Combination 1 required to have equal fat and protein contents between the sample for which a spectrum was to be estimated and the reference samples in the DHI database. Combination 2 also required an equal urea content. Combination 3 requested equal fat, protein, and lactose contents. Finally, combination 4 included all criteria. When more than one spectrum was found during the search, their average was the estimated spectrum for the query sample. Concretely, this study estimated missing spectra for 1,700 samples using 2,000,000 spectral DHI records. For assessing the effect of this spectral estimation on the prediction quality, FT-MIR equations were used to predict 11 phenotypes, selected as their quantification used different FT-MIR regions. They were related to the milk fat and mineral composition, lactoferrin content, quantity of eructed methane, body weight (BW), and dry matter intake. The accuracy between predictions obtained from actual and estimated spectra was evaluated by calculating the mean absolute error (MAE). The criteria in the fourth and second combinations were too strict to estimate a spectrum for most samples. Indeed, for many samples, no spectra with the same values for those matching criteria was found. The third match criteria combination had a poorer prediction performance for all studied traits and spectral absorptions than the first combination due to fewer matched samples available to compute the missing spectrum. By allowing a range for matching lactose content (±0.1 g/dL milk), we showed that this new combination increased the number of selected samples to compute missing spectra and predict better the infrared absorption at different wavenumbers, especially those related to the lactose quantification. The prediction performance was further improved by performing queries on the entire Walloon DHI spectral database (6,625,570 spectra), and it varied among the studied phenotypes. Without considering the traits used for the matching, the best predictions were obtained for the content of saturated fatty acids (MAE = 0.15 g/dL milk) and BW (MAE = 12.80 kg). Yet, the predictions for the unsaturated fatty acids were less accurate (MAE = 0.13 and 0.018 g/dL milk for monounsaturated and polyunsaturated fatty acids), likely because of the poorer predictions of spectral regions related to long-chain fatty acids. Similarly, poorer predictions were observed for the amount of methane eructed by dairy cows (MAE = 47.02 g/d), likely because it is not directly related to fat content or composition. Prediction accuracies for the remaining traits were also low. In conclusion, we observed that increasing the number of relevant matching criteria helps improve the quality of FT-MIR predicted phenotypes and the number of spectra used during the search. So, it would be of great interest to test in the future the suitability of the developed methodology with large-scale international spectral databases to improve the reliability of EBV from these FT-MIR-based phenotypes and the robustness of FT-MIR predictive models.

Single-step genome-wide association for selected milk fatty acids in Dual-Purpose Belgian Blue cows.

The aim of this study was to estimate genetic parameters and identify genomic regions associated with selected individual and groups of milk fatty acids (FA) predicted by milk mid-infrared spectrometry in Dual-Purpose Belgian Blue cows. The used data were 69,349 test-day records of milk yield, fat percentage, and protein percentage along with selected individual and groups FA of milk (g/dL milk) collected from 2007 to 2020 on 7,392 first-parity (40,903 test-day records), and 5,185 second-parity (28,446 test-day records) cows distributed in 104 herds in the Walloon Region of Belgium. Data of 28,466 SNPs, located on 29 Bos taurus autosomes (BTA), of 1,699 animals (639 males and 1,060 females) were used. Random regression test-day models were used to estimate genetic parameters through the Bayesian Gibbs sampling method. The SNP solutions were estimated using a single-step genomic best linear unbiased prediction approach. The proportion of genetic variance explained by each 25-SNP sliding window (with an average size of ~2 Mb) was calculated, and regions accounting for at least 1.0% of the total additive genetic variance were used to search for candidate genes. Average daily heritability estimated for the included milk FA traits ranged from 0.01 (C4:0) to 0.48 (C12:0) and 0.01 (C4:0) to 0.42 (C12:0) in the first and second parities, respectively. Genetic correlations found between milk yield and the studied individual milk FA, except for C18:0, C18:1 trans, C18:1 cis-9, were positive. The results showed that fat percentage and protein percentage were positively genetically correlated with all studied individual milk FA. Genome-wide association analyses identified 11 genomic regions distributed over 8 chromosomes [BTA1, BTA4, BTA10, BTA14 (4 regions), BTA19, BTA22, BTA24, and BTA26] associated with the studied FA traits, though those found on BTA14 partly overlapped. The genomic regions identified differed between parities and lactation stages. Although these differences in genomic regions detected may be due to the power of quantitative trait locus detection, it also suggests that candidate genes underlie the phenotypic expression of the studied traits may vary between parities and lactation stages. These findings increase our understanding about the genetic background of milk FA and can be used for the future implementation of genomic evaluation to improve milk FA profile in Dual-Purpose Belgian Blue cows.

Fourier transform mid-infrared milk screening to improve milk production and processing.

Milk mid-infrared spectrometry has been used for many years to quantify major milk compounds. Recently, much research has been conducted to extend the use of this technology to predict new, relevant phenotypes to assess the animals' welfare and the nutritional quality of milk, as well as its technological quality and environmental footprint. The transition from the research stage to field implementation is not easy, due to intrinsic and extrinsic constraints, but some developments can be considered to address these issues.

Can unsupervised learning methods applied to milk recording big data provide new insights into dairy cow health?

Among the dairy sector's current concerns, the assessment of global animal health status is a complex challenge. Its multidimensionality means that global monitoring tools are rarely considered. Instead, specific disease detection is often studied separately and, due to financial and ethical issues, uses small-scale data sets focusing on few biomarkers. Several studies have already been conducted using milk Fourier transform mid-infrared (FT-MIR) spectroscopy to detect mastitis and lameness or to quantify health-related biomarkers in milk or blood. Those studies are relevant but they focus mainly on one biomarker or disease. To solve this issue and the small-scale data set, in this study, we proposed a holistic approach using big data obtained from milk recording, including milk yield, somatic cell count, and 27 FT-MIR-based predictors related to milk composition and animal health status. Using 740,454 records collected from 114,536 first-parity Holstein cows in southern Belgium, we performed repeated unsupervised learning algorithms based on Ward's agglomerative hierarchical clustering method to find potential interesting patterns. A divide-and-conquer approach was used to overcome the limitation of computational resources in clustering a relatively large data set. Five groups of records were identified. Differences observed in the fourth group suggested a relationship to metabolic disorders. The fifth group seemed to be related to mastitis. In a second step, we performed a partial least squares discriminant analysis (PLS-DA) to predict the probability of belonging to those specific groups for the entire data set. The obtained global accuracy was 0.77 and the balanced accuracy (i.e., the mean between sensitivity and specificity) of discriminating the fourth and fifth groups was 0.88 and 0.96, respectively. Then, a validation of the interpretation of those groups was performed using 204 milk and blood reference records. The predicted probability associated with the metabolic disorders issue had significant correlations of 0.54 with blood ß-hydroxybutyrate, 0.44 with blood nonesterified fatty acids, -0.32 with blood glucose, -0.23 with milk glucose-6-phosphate, and 0.38 with milk isocitrate. In contrast, the predicted probability of belonging to the mastitis group had correlations of 0.69 with milk lactate dehydrogenase, 0.46 with milk N-acetyl-ß-d-glucosaminidase, -0.18 with milk free glucose, and 0.16 with milk glucose-6-phosphate. Consequently, these results suggest that the obtained quantitative traits indirectly reflect some of the main health disorders in dairy farming and could be used to monitor dairy cows on a large scale. By using unsupervised learning on large-scale milk recording data and then validating the pattern using reference laboratory measures, we propose a new approach to quickly assess dairy cow health status.

Identification of chronic stress biomarkers in dairy cows.

Stress in dairy herds can occur from multiple sources. When stress becomes chronic because of a long duration and inability of animals to adapt, it is likely to deeply affect the emotional state, health, immunity, fertility and milk production of cows. While assessing chronic stress in herds would be beneficial, no real consensus has emerged from the literature regarding the indicators of interest. The goal of this study was to compare and evaluate potential biomarkers for chronic stress after inducing stress over a 4-week period through severe overstocking, restricted access to feed and isolated unusual events. A total of 30 cows were involved in the experiment and two similar groups were constituted. Over a 4-week period, the 15 cows of the stress group were housed in overstocked conditions, with 4.6 m2 per cow, including resting and feeding areas. In this area, only seven individual places at the feeding area were available for the 15 cows to generate competition for feed access. Twice during the trial and over a period of 2 h, an additional stress was induced by moving cows to an unfamiliar barn and diffusion of stressing noises (dog barking). Meanwhile, the 15 cows of the control group stayed in the original barn, with more than 10 m2 per cow and more individual places at the feeding area than cow number. On a weekly basis, several variables considered as potential biomarkers for chronic stress were recorded. Collected data were analysed using single trait linear repeated mixed models. No differences were observed regarding milk yield, BW of cows or body condition score but the milk loss was more pronounced in the stress group. The activity was more heterogeneous and the rumination of cows was lower in the stress group. The heart rate was lower in the stress group and showed more heterogeneity at the end of the stress period. No differences were observed regarding salivary cortisol, blood glucose, ß-endorphin, thyroxine and leucocyte profile. A higher level of hair cortisol and blood fructosamine were observed in the stress group at the end of the stress period. Regarding the practical use of the highlighted biomarkers, milk loss may be an effective and easy way to detect general problems, including stress. The blood fructosamine and the hair cortisol concentrations are promising indicators to assess chronic stress in commercial farms.

Genetic analysis of milk urea concentration and its genetic relationship with selected traits of interest in dairy cows.

The aim of this study was to estimate genetic parameters of milk urea concentration (MU) and its genetic correlations with milk production traits, longevity, and functional traits in the first 3 parities in dairy cows. The edited data set consisted in 9,107,349 MU test-day records from the first 3 parities of 560,739 cows in 2,356 herds collected during the years 1994 to 2020. To estimate the genetic parameters of MU, data of 109 randomly selected herds, with a total of 770,016 MU test-day records, were used. Genetic parameters and estimated breeding values were estimated using a multiple-trait (parity) random regression model. Herd-test-day, age-year-season of calving, and days in milk classes (every 5 d as a class) were used as fixed effects, whereas effects of herd-year of calving, permanent environment, and animal were modeled using random regressions and Legendre polynomials of order 2. The average daily heritability and repeatability of MU during days in milk 5 to 365 in the first 3 parities were 0.19, 0.22, 0.20, and 0.48, 0.48, 0.47, respectively. The mean genetic correlation estimated among MU in the first 3 parities ranged from 0.96 to 0.97. The average daily estimated breeding values for MU of the selected bulls (n = 1,900) ranged from -9.09 to 7.37 mg/dL. In the last 10 yr, the genetic trend of MU has gradually increased. The genetic correlation between MU and 11 traits of interest ranged from -0.28 (milk yield) to 0.28 (somatic cell score). The findings of this study can be used as the first step for development of a routine genetic evaluation for MU and its inclusion into the genetic selection program in the Walloon Region of Belgium.

Large-scale phenotyping in dairy sector using milk MIR spectra: Key factors affecting the quality of predictions.

Methods and technologies enabling the estimation at large scale of important traits for the dairy sector are of great interest. Those phenotypes are necessary to improve herd management, animal genetic evaluation, and milk quality control. In the recent years, the research was very active to predict new phenotypes from the mid-infrared (MIR) analysis of milk. Models were developed to predict phenotypes such as fine milk composition, milk technological properties or traits related to cow health, fertility and environmental impact. Most of models were developed within research contexts and often not designed for routine use. The implementation of models at a large scale to predict new traits of interest brings new challenges as the factors influencing the robustness of models are poorly documented. The first objective of this work is to highlight the impact on prediction accuracy of factors such as the variability of the spectral and reference data, the spectral regions used and the complexity of models. The second objective is to emphasize methods and indicators to evaluate the quality of models and the quality of predictions generated under routine conditions. The last objective is to outline the issues and the solutions linked with the use and transfer of models on large number of instruments. Based on partial least square regression and 10 datasets including milk MIR spectra and reference quantitative values for 57 traits of interest, the impact of the different factors is illustrated by evaluating the influence on the validation root mean square error of prediction (RMSEP). In the displayed examples, all factors, when well set up, increase the quality of predictions, with an improvement of the RMSEP ranging from 12% to 43%. This work also aims to underline the need for and the complementarity between different validation procedures, statistical parameters and quality assurance methods. Finally, when using and transferring models, the impact of the spectral standardization on the prediction reproducibility is highlighted with an improvement up to 86% with the tested models, and the monitoring of individual spectrometer stability over time appears essential. This list inspired from our experience is of course not exhaustive. The displayed results are only examples and not general rules and other aspects play a role in the quality of final predictions. However, this work highlights good practices, methods and indicators to increase and evaluate quality of phenotypes predicted at a large scale. The results obtained argue for the development of guidelines at international levels, as well as international collaborations in order to constitute large and robust datasets and enable the use of models in routine conditions.

A comparison of 4 different machine learning algorithms to predict lactoferrin content in bovine milk from mid-infrared spectra.

Lactoferrin (LF) is a glycoprotein naturally present in milk. Its content varies throughout lactation, but also with mastitis; therefore it is a potential additional indicator of udder health beyond somatic cell count. Condequently, there is an interest in quantifying this biomolecule routinely. First prediction equations proposed in the literature to predict the content in milk using milk mid-infrared spectrometry were built using partial least square regression (PLSR) due to the limited size of the data set. Thanks to a large data set, the current study aimed to test 4 different machine learning algorithms using a large data set comprising 6,619 records collected across different herds, breeds, and countries. The first algorithm was a PLSR, as used in past investigations. The second and third algorithms used partial least square (PLS) factors combined with a linear and polynomial support vector regression (PLS + SVR). The fourth algorithm also used PLS factors, but included in an artificial neural network with 1 hidden layer (PLS + ANN). The training and validation sets comprised 5,541 and 836 records, respectively. Even if the calibration prediction performances were the best for PLS + polynomial SVR, their validation prediction performances were the worst. The 3 other algorithms had similar validation performances. Indeed, the validation root mean squared error (RMSE) ranged between 162.17 and 166.75 mg/L of milk. However, the lower standard deviation of cross-validation RMSE and the better normality of the residual distribution observed for PLS + ANN suggest that this modeling was more suitable to predict the LF content in milk from milk mid-infrared spectra (R2v = 0.60 and validation RMSE = 162.17 mg/L of milk). This PLS +ANN model was then applied to almost 6 million spectral records. The predicted LF showed the expected relationships with milk yield, somatic cell score, somatic cell count, and stage of lactation. The model tended to underestimate high LF values (higher than 600 mg/L of milk). However, if the prediction threshold was set to 500 mg/L, 82% of samples from the validation having a content of LF higher than 600 mg/L were detected. Future research should aim to increase the number of those extremely high LF records in the calibration set.

Short communication: Potential prediction of vitamin B12 concentration based on mid-infrared spectral data using Holstein Dairy Herd Improvement milk samples.

The purpose of this study was (1) to predict the quantitative concentration of vitamin B12 in milk using mid-infrared (MIR) spectrometry, and (2) to evaluate the potential of MIR spectra to discriminate different clusters of records based on their B12 concentration. Milk samples were collected from 4,340 Holstein cows between 3 and 592 d in milk and located in 100 herds. Samples were taken using in-line milk meters and divided into 2 aliquots: one for MIR spectrometry and the other for B12 concentration reference analyses by radioassay. Analyses were performed on 311 selected spectral wavelengths. A partial least squares regression model was built to quantify B12 concentration. Discriminant analysis was used to isolate B12 concentration clusters. A B12 concentration threshold was set at 442 ng/dL, because this represents the cutoff value for a 250-mL glass of milk to fulfill 46% of the daily vitamin B12 recommended dietary allowance for individuals 14 yr or older. For each analysis, records coming from two-thirds of herds were used to calibrate prediction equations, and the remaining records (one-third of herds for validation) were used to assess the prediction performance. In the case of discriminant analysis, validation sets were divided into evaluation sets (one-third of herds) to obtain alternate probability cutoffs and in test sets (two-thirds of herds) to validate equations. Spectral and B12 concentration outliers were identified by calculating standardized Mahalanobis distance and with a residual analysis, respectively (n = 3,154). Regarding quantitative B12 concentration, cross-validation and validation coefficients of determination averaged 0.51 and 0.46, respectively, which are relatively low, which would limit the potential use of the developed quantitative equations. In addition, root mean square errors of prediction of cross validation and validation sets averaged 88.9 and 94.7 ng/dL, respectively. Area under the receiver operating characteristic curve of test sets averaged 0.81 based on the 442 ng/dL threshold, which could be considered to represent good accuracy of classification. However, the false discovery rate averaged 36%. In summary, models predicting quantitative B12 concentration had low cross-validation and validation coefficients of determination, limiting their use, but the proposed discriminant models could be used to identify milk samples with naturally high B12.

Predicting milk mid-infrared spectra from first-parity Holstein cows using a test-day mixed model with the perspective of herd management.

The use of test-day models to model milk mid-infrared (MIR) spectra for genetic purposes has already been explored; however, little attention has been given to their use to predict milk MIR spectra for management purposes. The aim of this paper was to study the ability of a test-day mixed model to predict milk MIR spectra for management purposes. A data set containing 467,496 test-day observations from 53,781 Holstein dairy cows in first lactation was used for model building. Principal component analysis was implemented on the selected 311 MIR spectral wavenumbers to reduce the number of traits for modeling; 12 principal components (PC) were retained, explaining approximately 96% of the total spectral variation. Each of the retained PC was modeled using a single trait test-day mixed model. The model solutions were used to compute the predicted scores of each PC, followed by a back-transformation to obtain the 311 predicted MIR spectral wavenumbers. Four new data sets, containing altogether 122,032 records, were used to test the ability of the model to predict milk MIR spectra in 4 distinct scenarios with different levels of information about the cows. The average correlation between observed and predicted values of each spectral wavenumber was 0.85 for the modeling data set and ranged from 0.36 to 0.62 for the scenarios. Correlations between milk fat, protein, and lactose contents predicted from the observed spectra and from the modeled spectra ranged from 0.83 to 0.89 for the modeling set and from 0.32 to 0.73 for the scenarios. Our results demonstrated a moderate but promising ability to predict milk MIR spectra using a test-day mixed model. Current and future MIR traits prediction equations could be applied on the modeled spectra to predict all MIR traits in different situations instead of developing one test-day model separately for each trait. Modeling MIR spectra would benefit farmers for cow and herd management, for instance through prediction of future records or comparison between observed and expected wavenumbers or MIR traits for the detection of health and management problems. Potential resulting tools could be incorporated into milk recording systems.

Diagnosing the pregnancy status of dairy cows: How useful is milk mid-infrared spectroscopy?

Pregnancy diagnosis is an essential part of successful breeding programs on dairy farms. Milk composition alters with pregnancy, and this is well documented. Fourier-transform mid-infrared (MIR) spectroscopy is a rapid and cost-effective method for providing milk spectra that reflect the detailed composition of milk samples. Therefore, the aim of this study was to assess the ability of MIR spectroscopy to predict the pregnancy status of dairy cows. The MIR spectra and insemination records were available from 8,064 Holstein cows of 19 commercial dairy farms in Australia. Three strategies were studied to classify cows as open or pregnant using partial least squares discriminant analysis models with random cow-independent 10-fold cross-validation and external validation on a cow-independent test set. The first strategy considered 6,754 MIR spectra after insemination used as independent variables in the model. The results showed little ability to detect the pregnancy status as the area under the receiver operating characteristic curve was 0.63 and 0.65 for cross-validation and testing, respectively. The second strategy, involving 1,664 records, aimed to reduce noise in the MIR spectra used as predictors by subtracting a spectrum before insemination (i.e., open spectrum) from the spectrum after insemination. The accuracy was comparable with the first approach, showing no superiority of the method. Given the limited results for these models when using combined data from all stages after insemination, the third strategy explored separate models at 7 stages after insemination comprising 348 to 1,566 records each (i.e., progressively greater gestation) with single MIR spectra after insemination as predictors. The models developed using data recorded after 150 d of pregnancy showed promising prediction accuracy with the average value of area under the receiver operating characteristic curve of 0.78 and 0.76 obtained through cross-validation and testing, respectively. If this can be confirmed on a larger data set and extended to somewhat earlier stages after insemination, the model could be used as a complementary tool to detect fetal abortion.

Short communication: Development of an equation for estimating methane emissions of dairy cows from milk Fourier transform mid-infrared spectra by using reference data obtained exclusively from respiration chambers.

Evaluation and mitigation of enteric methane (CH4) emissions from ruminant livestock, in particular from dairy cows, have acquired global importance for sustainable, climate-smart cattle production. Based on CH4 reference measurements obtained with the SF6 tracer technique to determine ruminal CH4 production, a current equation permits evaluation of individual daily CH4 emissions of dairy cows based on milk Fourier transform mid-infrared (FT-MIR) spectra. However, the respiration chamber (RC) technique is considered to be more accurate than SF6 to measure CH4 production from cattle. This study aimed to develop an equation that allows estimating CH4 emissions of lactating cows recorded in an RC from corresponding milk FT-MIR spectra and to challenge its robustness and relevance through validation processes and its application on a milk spectral database. This would permit confirming the conclusions drawn with the existing equation based on SF6 reference measurements regarding the potential to estimate daily CH4 emissions of dairy cows from milk FT-MIR spectra. A total of 584 RC reference CH4 measurements (mean ± standard deviation of 400 ± 72 g of CH4/d) and corresponding standardized milk mid-infrared spectra were obtained from 148 individual lactating cows between 7 and 321 d in milk in 5 European countries (Germany, Switzerland, Denmark, France, and Northern Ireland). The developed equation based on RC measurements showed calibration and cross-validation coefficients of determination of 0.65 and 0.57, respectively, which is lower than those obtained earlier by the equation based on 532 SF6 measurements (0.74 and 0.70, respectively). This means that the RC-based model is unable to explain the variability observed in the corresponding reference data as well as the SF6-based model. The standard errors of calibration and cross-validation were lower for the RC model (43 and 47 g/d vs. 66 and 70 g/d for the SF6 version, respectively), indicating that the model based on RC data was closer to actual values. The root mean squared error (RMSE) of calibration of 42 g/d represents only 10% of the overall daily CH4 production, which is 23 g/d lower than the RMSE for the SF6-based equation. During the external validation step an RMSE of 62 g/d was observed. When the RC equation was applied to a standardized spectral database of milk recordings collected in the Walloon region of Belgium between January 2012 and December 2017 (1,515,137 spectra from 132,658 lactating cows in 1,176 different herds), an average ± standard deviation of 446 ± 51 g of CH4/d was estimated, which is consistent with the range of the values measured using both RC and SF6 techniques. This study confirmed that milk FT-MIR spectra could be used as a potential proxy to estimate daily CH4 emissions from dairy cows provided that the variability to predict is covered by the model.

The feeding system impacts relationships between calving interval and economic results of dairy farms.

The calving interval (CI) can potentially impact the economic results of dairy farms. This study highlighted the most profitable CI and innovated by describing this optimum as a function of the feeding system of the farm. On-farm data were used to represent real farm conditions. A total of 1832 accounts of farms recorded from 2007 to 2014 provided economic, technical and feeding information per herd and per year. A multiple correspondence analysis created four feeding groups: extensive, low intensive, intensive and very intensive herds. The gross margin and some of its components were corrected to account for the effect of factors external to the farm, such as the market, biological status, etc. Then the corrected gross margin (cGMc) and its components were modelled by CI parameters in each feeding system by use of GLM. The relationship between cGMc and the proportion of cows with CI<380 days in each feeding group showed that keeping most of the cows in the herd with CI near to 1 year was not profitable for most farms (for the very intensive farms there was no effect of the proportion). Moreover, a low proportion of cows (0% to 20%) with a near-to-1-year CI was not profitable for the extensive and low intensive farms. Extending the proportion of cows with CI beyond 459 days until 635 days (i.e. data limitation) caused no significant economic loss for the extensive and low intensive farms, but was not profitable for the intensive and very intensive farms. Variations of the milk and feeding components explained mainly these significant differences of gross margin. A link between the feeding system and persistency, perceptible in the milk production and CI shown by the herd, could explain the different relationships observed between the extent of CI and the economic results in the feeding groups. This herd-level study tended to show different economic optima of CI as a function of the feeding system. A cow-level study would specify these tendencies to give CI objectives to dairy breeders as a function of their farm characteristics.

Relationships between milk mid-IR predicted gastro-enteric methane production and the technical and financial performance of commercial dairy herds.

Considering economic and environmental issues is important in ensuring the sustainability of dairy farms. The objective of this study was to investigate univariate relationships between lactating dairy cow gastro-enteric methane (CH4) production predicted from milk mid-IR (MIR) spectra and technico-economic variables by the use of large scale and on-farm data. A total of 525 697 individual CH4 predictions from milk MIR spectra (MIR-CH4 (g/day)) of milk samples collected on 206 farms during the Walloon milk recording scheme were used to create a MIR-CH4 prediction for each herd and year (HYMIR-CH4). These predictions were merged with dairy herd accounting data. This allowed a simultaneous study of HYMIR-CH4 and 42 technical and economic variables for 1024 herd and year records from 2007 to 2014. Pearson correlation coefficients (r) were used to assess significant relationships (P<0.05). Low HYMIR-CH4 was significantly associated with, amongst others, lower fat and protein corrected milk (FPCM) yield (r=0.18), lower milk fat and protein content (r=0.38 and 0.33, respectively), lower quantity of milk produced from forages (r=0.12) and suboptimal reproduction and health performance (e.g. longer calving interval (r=-0.21) and higher culling rate (r=-0.15)). Concerning economic results, low HYMIR-CH4 was significantly associated with lower gross margin per cow (r=0.19) and per litre FPCM (r=0.09). To conclude, this study suggested that low lactating dairy cow gastro-enteric CH4 production tended to be associated with more extensive or suboptimal management practices, which could lead to lower profitability. The observed low correlations suggest complex interactions between variables due to the use of on-farm data with large variability in technical and management practices.

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

An increasing number of models are being developed to provide information from milk Fourier transform mid-infrared (FT-MIR) spectra on fine milk composition, technological properties of milk, or even cows' physiological status. In this context, and to take advantage of these existing models, the purpose of this work was to evaluate whether a spectral standardization method can enable the use of multiple equations within a network of different FT-MIR spectrometers. The piecewise direct standardization method was used, matching "slave" instruments to a common reference, the "master." The effect of standardization on network reproducibility was assessed on 66 instruments from 3 different brands by comparing the spectral variability of the slaves and the master with and without standardization. With standardization, the global Mahalanobis distance from the slave spectra to the master spectra was reduced on average from 2,655.9 to 14.3, representing a significant reduction of noninformative spectral variability. The transfer of models from instrument to instrument was tested using 3 FT-MIR models predicting (1) the quantity of daily methane emitted by dairy cows, (2) the concentration of polyunsaturated fatty acids in milk, and (3) the fresh cheese yield. The differences, in terms of root mean squared error, between master predictions and slave predictions were reduced after standardization on average from 103 to 17 g/d, from 0.0315 to 0.0045 g/100 mL of milk, and from 2.55 to 0.49 g of curd/100 g of milk, respectively. For all the models, standard deviations of predictions among all the instruments were also reduced by 5.11 times for methane, 5.01 times for polyunsaturated fatty acids, and 7.05 times for fresh cheese yield, showing an improvement of prediction reproducibility within the network. Regarding the results obtained, spectral standardization allows the transfer and use of multiple models on all instruments as well as the improvement of spectral and prediction reproducibility within the network. The method makes the models universal, thereby offering opportunities for data exchange and the creation and use of common robust models at an international level to provide more information to the dairy sector from direct analysis of milk.

Genetic parameters of mid-infrared methane predictions and their relationships with milk production traits in Holstein cattle.

Many countries have pledged to reduce greenhouse gases. In this context, the dairy sector is one of the identified sectors to adapt production circumstances to address socio-environmental constraints due to its large carbon footprint related to CH4 emission. This study aimed mainly to estimate (1) the genetic parameters of 2 milk mid-infrared-based CH4 proxies [predicted daily CH4 emission (PME, g/d), and log-transformed predicted CH4 intensity (LMI)] and (2) their genetic correlations with milk production traits [milk (MY), fat (FY), and protein (PY) yields] from first- and second-parity Holstein cows. A total of 336,126 and 231,400 mid-infrared CH4 phenotypes were collected from 56,957 and 34,992 first- and second-parity cows, respectively. The PME increased from the first to the second lactation (433 vs. 453 g/d) and the LMI decreased (2.93 vs. 2.86). We used 20 bivariate random regression test-day models to estimate the variance components. Moderate heritability values were observed for both CH4 traits, and those values decreased slightly from the first to the second lactation (0.25 ± 0.01 and 0.22 ± 0.01 for PME; 0.18 ± 0.01 and 0.17 ± 0.02 for LMI). Lactation phenotypic and genetic correlations were negative between PME and MY in both first and second lactations (-0.07 vs. -0.07 and -0.19 vs. -0.24, respectively). More close scrutiny revealed that relative increase of PME was lower with high MY levels even reverting to decrease, and therefore explaining the negative correlations, indicating that higher producing cows could be a mitigation option for CH4 emission. The PME phenotypic correlations were almost equal to 0 with FY and PY for both lactations. However, the genetic correlations between PME and FY were slightly positive (0.11 and 0.12), whereas with PY the correlations were slightly negative (-0.05 and -0.04). Both phenotypic and genetic correlations between LMI and MY or PY or FY were always relatively highly negative (from -0.21 to -0.88). As the genetic correlations between PME and LMI were strong (0.71 and 0.72 in first and second lactation), the selection of one trait would also strongly influence the other trait. However, in animal breeding context, PME, as a direct quantity CH4 proxy, would be preferred to LMI, which is a ratio trait of PME with a trait already in the index. The range of PME sire estimated breeding values were 22.1 and 29.41 kg per lactation in first and second parity, respectively. Further studies must be conducted to evaluate the effect of the introduction of PME in a selection index on the other traits already included in this index, such as, for instance, fertility or longevity.

Invited review: Large-scale indirect measurements for enteric methane emissions in dairy cattle: A review of proxies and their potential for use in management and breeding decisions.

Efforts to reduce the carbon footprint of milk production through selection and management of low-emitting cows require accurate and large-scale measurements of methane (CH4) emissions from individual cows. Several techniques have been developed to measure CH4 in a research setting but most are not suitable for large-scale recording on farm. Several groups have explored proxies (i.e., indicators or indirect traits) for CH4; ideally these should be accurate, inexpensive, and amenable to being recorded individually on a large scale. This review (1) systematically describes the biological basis of current potential CH4 proxies for dairy cattle; (2) assesses the accuracy and predictive power of single proxies and determines the added value of combining proxies; (3) provides a critical evaluation of the relative merit of the main proxies in terms of their simplicity, cost, accuracy, invasiveness, and throughput; and (4) discusses their suitability as selection traits. The proxies range from simple and low-cost measurements such as body weight and high-throughput milk mid-infrared spectroscopy (MIR) to more challenging measures such as rumen morphology, rumen metabolites, or microbiome profiling. Proxies based on rumen samples are generally poor to moderately accurate predictors of CH4, and are costly and difficult to measure routinely on-farm. Proxies related to body weight or milk yield and composition, on the other hand, are relatively simple, inexpensive, and high throughput, and are easier to implement in practice. In particular, milk MIR, along with covariates such as lactation stage, are a promising option for prediction of CH4 emission in dairy cows. No single proxy was found to accurately predict CH4, and combinations of 2 or more proxies are likely to be a better solution. Combining proxies can increase the accuracy of predictions by 15 to 35%, mainly because different proxies describe independent sources of variation in CH4 and one proxy can correct for shortcomings in the other(s). The most important applications of CH4 proxies are in dairy cattle management and breeding for lower environmental impact. When breeding for traits of lower environmental impact, single or multiple proxies can be used as indirect criteria for the breeding objective, but care should be taken to avoid unfavorable correlated responses. Finally, although combinations of proxies appear to provide the most accurate estimates of CH4, the greatest limitation today is the lack of robustness in their general applicability. Future efforts should therefore be directed toward developing combinations of proxies that are robust and applicable across diverse production systems and environments.

Invited review: Phenotypes to genetically reduce greenhouse gas emissions in dairying.

Phenotypes have been reviewed to select for lower-emitting animals in order to decrease the environmental footprint of dairy cattle products. This includes direct selection for breath measurements, as well as indirect selection via indicator traits such as feed intake, milk spectral data, and rumen microbial communities. Many of these traits are expensive or difficult to record, or both, but with genomic selection, inclusion of methane emission as a breeding goal trait is feasible, even with a limited number of registrations. At present, methane emission is not included among breeding goals for dairy cattle worldwide. There is no incentive to include enteric methane in breeding goals, although global warming and the release of greenhouse gases is a much-debated political topic. However, if selection for reduced methane emission became a reality, there would be limited consensus as to which phenotype to select for: methane in liters per day or grams per day, methane in liters per kilogram of energy-corrected milk or dry matter intake, or a residual methane phenotype, where methane production is corrected for milk production and the weight of the cow. We have reviewed the advantages and disadvantages of these traits, and discuss the methods for selection and consequences for these phenotypes.

Changes throughout lactation in phenotypic and genetic correlations between methane emissions and milk fatty acid contents predicted from milk mid-infrared spectra.

The aim of this study was to estimate phenotypic and genetic correlations between methane production (Mp) and milk fatty acid contents of first-parity Walloon Holstein cows throughout lactation. Calibration equations predicting daily Mp (g/d) and milk fatty acid contents (g/100 dL of milk) were applied on milk mid-infrared spectra related to Walloon milk recording. A total of 241,236 predictions of Mp and milk fatty acids were used. These data were collected between 5 and 305 d in milk in 33,555 first-parity Holstein cows from 626 herds. Pedigree data included 109,975 animals. Bivariate (i.e., Mp and a fatty acid trait) random regression test-day models were developed to estimate phenotypic and genetic parameters of Mp and milk fatty acids. Individual short-chain fatty acids (SCFA) and groups of saturated fatty acids, SCFA, and medium-chain fatty acids showed positive phenotypic and genetic correlations with Mp (from 0.10 to 0.16 and from 0.23 to 0.30 for phenotypic and genetic correlations, respectively), whereas individual long-chain fatty acids (LCFA), and groups of LCFA, monounsaturated fatty acids, and unsaturated fatty acids showed null to positive phenotypic and genetic correlations with Mp (from -0.03 to 0.13 and from -0.02 to 0.32 for phenotypic and genetic correlations, respectively). However, these correlations changed throughout lactation. First, de novo individual and group fatty acids (i.e., C4:0, C6:0, C8:0, C10:0, C12:0, C14:0, SCFA group) showed low phenotypic or genetic correlations (or both) in early lactation and higher at the end of lactation. In contrast, phenotypic and genetic correlations between Mp and C16:0, which could be de novo synthetized or derived from blood lipids, were more stable during lactation. This fatty acid is the most abundant fatty acid of the saturated fatty acid and medium-chain fatty acid groups of which correlations with Mp showed the same pattern across lactation. Phenotypic and genetic correlations between Mp and C17:0 and C18:0 were low in early lactation and increased afterward. Phenotypic and genetic correlations between Mp and C18:1 cis-9 originating from the blood lipids were negative in early lactation and increased afterward to become null from 18 wk until the end of lactation. Correlations between Mp and groups of LCFA, monounsaturated fatty acids, and unsaturated fatty acids showed a similar or intermediate pattern across lactation compared with fatty acids that compose them. Finally, these results indicate that correlations between Mp and milk fatty acids vary following lactation stage of the cow, a fact still often ignored when trying to predict Mp from milk fatty acid profile.