Recent studies suggest that variations in milk yield during lactation can serve as a metric for assessing individual resilience, with minimum fluctuations reflecting greater resilience in cows. With automatic milking systems, multiple records can be obtained from many cows on a daily basis. Given that animals are continually exposed to a range of stressors throughout their productive life, identifying resilience traits is crucial for preserving both their productive and reproductive capacities. We studied two resilience indicators: natural log-transformed variance (ln(σ2)) and lag-1 autocorrelation (rlag) of the daily deviations in milk yield recorded from the automatic milking system, with average daily milk yield incorporated as an additional trait to assess the correlations between the resilience indicators and milk yield. The resilience indicators were derived from three fitted lactation curves: the Wilmink curve, the penalised quadratic spline and the penalised cubic spline. The data consisted of 18,549,535 daily milk records of 37,118 Fleckvieh, 7322 Holstein-Friesian and 2784 Brown Swiss cows in 1070 Austrian dairy herds. The data were collected for lactation periods of up to 350 days across multiple parities. Additionally, the resilience indicators were calculated based on days 11-50 of lactation, as the transition period and early lactation are crucial moments for the cows. Genetic analyses were conducted using univariate and bivariate analyses to estimate heritability (h2) and genetic correlations between the traits. Estimated breeding values of the indicator traits obtained from the genetic analysis were correlated with routinely estimated breeding values and official indexes of the cows, including dairy, health, functional and conformation traits. The resilience indicators were scaled to negatives (-ln(σ2) and -rlag) to compute these correlations, so that positive values indicate better resilience. The h2 ranged between 0.09 and 0.20 for ln(σ2) and between 0.04 and 0.06 for rlag, with the highest estimates observed when the penalised cubic spline was used. Generally, the h2 decreased when only the early lactation period was considered. Genetic correlations of the resilience indicators with existing traits were low to moderate, with stronger correlations observed with -ln(σ2) than in -rlag, particularly when estimated with the dairy traits. The negative correlations with milk yield, ranging from -0.13 to -0.45 for -ln(σ2) and -0.03 to -0.23 for -rlag suggest that selecting for the resilience traits may result in a reduction in milk yield. This study demonstrated that deviations derived from fluctuations in lactation curves are heritable. Among the two resilience indicators evaluated, -ln(σ2) emerged as the more suitable trait for selection, considering its h2 and positive correlation to health and functional traits. These findings underscore the potential for incorporating resilience traits into breeding programs while emphasising the need to balance improving resilience with milk production, considering the probable risk of reduced milk production in cows.

Accurate estimation of methane (CH4) and carbon dioxide (CO2) emissions requires precise sampling methods due to the significant variability of these gases. This study, part of the project breed4green in Austria, investigated the variability of CH4 and CO2 and minimum number of spot samples required to accurately estimate CH4 and CO2 emissions from dairy cows on commercial farms. A total of 51,895 spot samples from 451 cows on 9 farms was analyzed; these samples were obtained using the GreenFeed system. Each farm participated in 2 GreenFeed measurement periods (period 1 and 2), each lasting 6 wk and separated by 3 to 6 mo. A total of 298 cows, each with at least 70 spot samples in either period 1 or period 2, were analyzed to determine the minimum number of spot samples. Two methods were used. In the first, forward sampling, subsets of increasing spot samples (the first 1, 10, 20, 30, 40, 50, 60, and 70 measurements per cow) were compared with the complete set of available measurements for the cows. In the second method, random sampling, the same number of spot samples was selected randomly from all available visits for each cow and this procedure was repeated 100 times to account for sampling variability. The minimum number of spot samples required was determined when the Pearson correlation between the arithmetic mean of the subset and that of all available measurements reached a value of 0.90. Mean CH4 and CO2 emissions were 430 g/d and 13,309 g/d, respectively. The CV for diurnal variability ranged from 4.6% to 12.0% for CH4 and from 2.7% to 6.9% for CO2, indicating low to moderate diurnal variability. Day-to-day variability was low, with CV between 3.7% and 8.7% for CH4 and between 2.5% and 5.5% for CO2. Substantial within-cow variation was observed for CH4 (mean CV of 22.1% in period 1 and 22.9% in period 2), whereas CO2 was more stable (mean CV of 11.8% in period 1 and 11.5% in period 2). When using forward sampling, at least 19 spot samples were required for accurate CH4 estimates and only 13 for CO2, with a mean of 6.5 and 4.3 d needed to obtain these measurements, respectively. When using random sampling, only 8 spot samples for CH4 and 5 spot samples for CO2 were sufficient to meet the accuracy criteria. For future analyses of this dataset, we will rely on the results of forward sampling. Forward sampling is based on the actual sequence of visits made by each cow to the GreenFeed and therefore offers a more conservative and realistic approach compared with random sampling. Overall, the results highlight the dynamic nature of CH4 emissions, which are influenced by factors such as feeding patterns and rumen fluctuations and suggest that more spot samples are needed for reliable estimates of CH4 than for CO2.

Milk mid-infrared (MIR) spectroscopy offers a rapid and cost-effective method for quantifying milk components and has shown potential for predicting physiological and environmental traits in dairy cattle, including methane (CH4) emissions. However, milk MIR-based CH4 predictions remain a black-box, as the biological mechanisms linking milk MIR spectra to CH4 emissions are not understood. This study aimed to predict CH4 emissions in dairy cows, measured using the GreenFeed system, based on milk MIR spectra and milk fatty acids (MFA). By comparing milk MIR-based models with those using MFA as predictors, we intended to disentangle the underlying biological signals associated with CH4 emissions. Data from 16 commercial Austrian dairy farms were available. All farms had a similar feeding system based on grass and corn silage, supplemented by varying amounts of concentrated feed and without grazing. Data included 807 records from 604 cows. Six partial least squares regression models were developed to predict CH4 emissions by using various combinations of predictor variables, including milk MIR spectra, MFA, major milk components, milk yield, days in milk and parity. A nested cross-validation framework was applied using 2 strategies: 5-fold cow-independent and leave-one-farm-out cross-validation. Mean CH4 emissions were 436 g/d. Positive correlations were observed between CH4 and de novo MFA throughout lactation, whereas preformed MFA showed negative associations, particularly in early lactation. These associations reflect signals associated with energy status. In 5-fold cow-independent cross-validation, the model based on milk MIR spectra showed moderate predictive accuracy (r = 0.52, R2 = 0.27, RMSE = 60 g/d), while the model based on milk MFA performed lower (r = 0.42, R2 = 0.18, RMSE = 63 g/d). However, combining MFA with fat%, protein% and lactose% slightly increased accuracy (r = 0.44, R2 = 0.20, RMSE = 62 g/d). The additional consideration of milk yield, days in milk and parity improved accuracy of MIR-based (r = 0.62, R2 = 0.39, RMSE = 55 g/d) and MFA-based models (r = 0.60, R2 = 0.36, RMSE = 56 g/d). The model including milk yield, days in milk and parity showed lower predictive performance (r = 0.49, R2 = 0.24, RMSE = 60 g/d), indicating that both MIR spectra and MFA together with fat%, protein% and lactose% provide additional information. Leave-one-farm-out cross-validation resulted in slightly lower predictive accuracies (MIR-based model r = 0.43, R2 = 0.22, RMSE = 63 g/d and MFA-based model: r = 0.41, R2 = 0.21, RMSE = 66 g/d). Models using all predictor traits performed best (MIR-based model r = 0.56, R2 = 0.33, RMSE = 58 g/d and MFA-based model: r = 0.58, R2 = 0.35, RMSE = 57 g/d). Correlations between predictions from the MIR- and MFA-based models including all predictors were high (r = 0.86), indicating a high agreement between the 2 modeling approaches. This study represents one of the first large-scale evaluations of milk MIR-based CH4 prediction models under commercial farm conditions. Milk MIR spectra and MFA explained only part of the variability in CH4 emissions, largely reflecting metabolic signals related to energy status. Correlations between CH4 emissions and MFA suggest that these models will favor cows with a negative energy balance during early lactation and with consistently lower de novo MFA synthesis throughout lactation. Further research is needed to gain more insight into the use of milk mid-infrared spectra for genetic selection to reduce CH4 emissions.

A negative energy balance in early lactation increases the susceptibility of dairy cows to metabolic and infectious diseases. Energy balance (EB) is therefore a valuable trait in both herd management and breeding strategies, to improve the health and efficiency of dairy cows. However, routine on-farm recording of energy intake is hardly feasible, necessitating suitable alternatives to determine EB or related traits. A potential alternative is prediction based on mid-infrared (MIR) spectral and test-day data from routine milk recording. Thus, the aim of the present study was to develop spectrometric prediction equations for EB and related traits, i.e., energy intake (EI) and dry matter intake (DMI), for Fleckvieh and Holstein Friesian dairy cows. A dataset was available comprising 64 988 daily observations of phenotypes including test-day variables and milk MIR spectra from 18 Fleckvieh and 71 Holstein Friesian cows, collected on a research farm between 2014 and 2021. Based on this dataset, quantitative prediction models were developed using different combinations of 212 selected first derivative MIR spectra and test-day variables by applying partial least squares regression analysis. An additional dataset was used for external validation by farm of the developed prediction equations, comprising 1 971 records on 16 Fleckvieh and 20 Holstein Friesian cows collected between 2017 and 2020 on another research farm. In addition to different validation scenarios, various effects, including breed, parity, and concentrate intake, were also evaluated for their impact on predictability of the traits considered. In general, prediction equations have shown to be most accurate when they included 212 MIR spectra along with parity and milk yield as predictors. The prediction equations provided moderate accuracies exhibiting correlation coefficients of 0.59 to 0.75 for EB, 0.63 to 0.71 for DMI, and 0.69 to 0.71 for EI, depending on the specific validation scenarios. The effects of breed, parity, and concentrate level showed differing impacts on the predictive capacity of the models for EB, DMI, and EI, with variations across traits. The results demonstrate potential for the generation of population-level phenotypes for EB, DMI, and EI based on routinely available MIR spectra and test-day variables. This approach would facilitate the routine recording of such indicators on a large scale for farm management and inclusion in genetic evaluation systems.

Milk lactose content (LC) physiologically decreases with parity order in dairy cows, but also after udder health inflammation(s) and in presence of elevated milk SCC in subclinical cases. Therefore, the progressive decrease in milk LC observed along cows' productive life can be attributed to a combination of factors that altogether impair the epithelial integrity, resulting in weaker tight junctions, e.g., physiological aging of epithelium, mechanical epithelial stress due to milking, and experienced clinical or subclinical mastitis. Mastitis is also known to affect the udder synthesis ability, so our intention through this study was to evaluate if there is a cumulative and lasting effect of mammary gland inflammation(s) on milk yield (MY) and LC. For this purpose, we used diagnoses of clinical mastitis and milk data of Austrian Fleckvieh cows to evaluate the effect of cumulative mastitis events on LC and MY. Only mastitis diagnoses recorded by trained veterinarians were used. Finally, we investigated if cumulative mastitis is a heritable trait and whether it is genetically correlated with either LC or MY. Estimates were obtained using univariate and bivariate linear animal models. A significant reduction in LC and MY was observed in cows that suffered from mastitis compared with those that did not experience udder inflammation. The h2 of cumulative mastitis is promising and much greater (0.09) than the h2 of the binary event itself (<= 0.03). The genetic correlations between cumulative mastitis with LC and MY were negative, suggesting that cows with a great genetic merit for MY and LC are expected to be more resistant to repeated inflammations and less recidivist. When we used number of lifetime SCC peaks (>= 200,000 or 400,000 cells/mL) to calculate cumulative inflammation events, h2 was even higher (up to 0.38), implying that subclinical mastitis also has a relevant negative impact on both LC and MY. Finally, the present study demonstrated how repeated mastitis events can permanently affect the mammary gland epithelial integrity and synthesis ability, and that the number of cumulative mastitis is a promising phenotype to be used in selection index in combination with other indicator traits toward more resistant and resilient mammary glands.

Lameness is an important health and welfare issue that causes considerable economic losses in dairy herds. The objective of this study was to investigate whether the hind feet position score (HFPS) can be used as an auxiliary trait for genetic evaluation of lameness. The HFPS is evaluated by visual scoring of the position of both the hind-digits to the mid-line of the cow's body. The higher the heel height of the lateral claw, the higher is the HFPS, and the higher is the risk for development of lameness. In total, 3,478 records from 1,064 Fleckvieh cows from 35 farms were obtained between September 1, 2021, and March 5, 2022. Data collection was carried out by the regional milk recording organizations. Hind feet position was scored visually by trained personnel during routine milk performance testing in the milking parlor using a 3-class scoring system: score 1 = 0° to <17° indicating a balanced heel height of both the medial and the lateral claw; score 2 = angle of 17° to 24°; score 3 = angle of >24°. After all cows had been milked, locomotion scoring was performed for each animal using a 5-class scoring system with locomotion scores ranging between 1 (normal) and 5 (severely lame). Using HFPS, sensitivity and specificity were 69.5% and 66.8%, respectively, for detecting lameness defined by locomotion score ≥2. For genetic analyses, a bivariate linear animal model was fitted with fixed effects of herd, parity, lactation stage, and classifier, and random effects of animal and permanent environment. Heritabilities for HFPS and locomotion score were 0.07 and 0.10, respectively, and the genetic correlation between the 2 traits studied was 0.80. These results suggest that the HFPS could be used for genetic evaluations to reduce lameness incidence in dairy cattle.

The deployment of diverse data-generating technologies in livestock farming holds the promise of early disease detection and improved animal well-being. In this paper, we combine routinely collected dairy farm and herd data with weather and high-frequency sensor data from 6 farms to predict new lameness events in various future periods, spanning from the following day to 3 wk. A Random Forest classifier, using input features selected by the Boruta algorithm, was used for the prediction task; effects of individual features were further assessed using partial dependence plots. We achieve precision scores of up to 93% when predicting lameness for the next 3 wk and when using information from the last 3 wk, combined with a balanced accuracy of 79%. Removing sensor data results has a tendency to reduce the precision for predictions, especially when using information from the last 1, 2, or 3 wk. Moving to a larger dataset (without sensor data) of 44 farms keeps the similar balanced accuracy but reduces precision by more than 30%, revealing a substantial a trade-off in model quality between false positives (false lameness alerts) and false negatives (missed lameness events). Sensor data holds promise to further improve the precision of these models, but can be partially compensated by high-resolution data from other systems, such as automated milking systems.

Claw health of Austrian dairy herds was evaluated using data collected in the project 'KlauenQ-Wohl' from 28,638 cattle in 526 dairy farms. We calculated the incidence of claw lesions and examined the relationships between cumulative incidence of claw lesions and lactation number, lactation month, housing type and breed. Claw health data were electronically documented by hoof trimmers from 2010 to 2019. Data were subjected to validity checks and hoof trimmers underwent inter-observer reliability testing. During the observation period of ten years, only 12.4 % of cows were hoof-trimmed at approximately and after 305 days, 28.3 % during their first 100 DIM and 59.3 % between their 101 and 305 DIM. The mean incidences during the ten-year-observation period were 59.2 % for heel horn erosion, 42.6 % for whiteline-lesions, 29.6 % for digital dermatitis (DD), 14.2 % for ulcers at all claw locations and 29.5 % for claw lesions that are always associated with pain and lameness ('alarm' lesions). Herd prevalence of DD in 2019 was 48.9 %. Cows in higher lactations had significantly higher incidences of concave dorsal walls, sole haemorrhages, sole ulcers, white-line-lesions and heel horn erosion, while heifers, cows in their first two lactations and cows around parturition showed significantly higher incidences of DD and interdigital phlegmon (foot rot). The mean incidence of all claw lesions was significantly (p<0.05) higher in cows kept in loose housing systems (85.3 %) than in cows kept in tie stalls (79.6 %). Fleckvieh cows had the highest overall incidence of all claw lesions of 89.5 %, followed by Holstein Friesian cows with 87.4 % and Brown Swiss cows with 72.1 %. 'Alarm' lesions (44.3 %) and DD (42.1 %) were most frequent in Holstein Friesian cows. To reduce the incidence of claw lesions, hoof trimming at dry-off and again around 40-60 DIM could be implemented, with significant benefits to claw health.

Data cleaning is a core process when it comes to using data from dairy sensor technologies. This article presents guidelines for sensor data cleaning with a specific focus on dairy herd management and breeding applications. Prior to any data cleaning steps, context and purpose of the data use must be considered. Recommendations for data cleaning are provided in five distinct steps: 1) validate the data merging process, 2) get to know the data, 3) check completeness of the data, 4) evaluate the plausibility of sensor measures and detect outliers, and 5) check for technology related noise. Whenever necessary, the recommendations are supported by examples of different sensor types (bolus, accelerometer) collected in an international project (D4Dairy) or supported by relevant literature. To ensure quality and reproducibility, data users are required to document their approach throughout the process. The target group for these guidelines are professionals involved in the process of collecting, managing, and analyzing sensor data from dairy herds. Providing guidelines for data cleaning could help to ensure that the data used for analysis is accurate, consistent, and reliable, ultimately leading to more informed management decisions and better breeding outcomes for dairy herds.

In the domain of precision livestock farming, the integration of diverse data sources is crucial for advancing sustainability and evaluating the implications of farm management practices on cow health. Addressing the challenge of data heterogeneity and management diversity, we propose a key-feature-based clustering method. This approach, merging knowledge-driven feature selection with unsupervised machine learning, enables the systematic investigation of management effects on cow health by forming distinct clusters for analysis. Utilizing data from 3,284 Austrian farms, including 80 features related to feeding, milking, housing, and technology systems, and health information for 56,000 cows, we show how this methodology can be applied to study the impact of technological systems on cow health resulting from the incidence of veterinary diagnoses. Our analysis successfully identified 14 distinct clusters, further divided into four main groups based on their level of technological integration in farm management: “SMART,” “TRADITIONAL,” “AMS (automatic milking system),” and “SENSOR.” We found that “SMART” farms, which integrate both AMS and sensor systems, exhibited a minimally higher disease risk for milk fever (OR 1.09) but lower risks for fertility disorders and udder diseases, indicating a general trend toward reduced disease risks. In contrast, farms with “TRADITIONAL” management, without AMS and sensor systems, showed the lowest risk for milk fever but the highest risk of udder disease (OR 1.12) and a minimally higher incidence of fertility disorders (OR 1.07). Furthermore, across all four groups, we observed that organic farming practices were associated with a reduced incidence of milk fever, udder issues, and particularly fertility diagnoses. However, the size of the effect varied by cluster, highlighting the complex and multifactorial nature of the relationship between farm management practices and disease risk. The study highlights the effectiveness of the key-feature-based clustering approach for high-dimensional data analyses aimed at comparing different management practices and exploring their complex relationships. The adaptable analytical framework of this approach makes it a promising tool for planning optimizing sustainable and efficient animal husbandry practices.

Milk analysis using mid-infrared spectroscopy (MIR) is a fast and inexpensive way of examining milk samples on a large scale for fat, protein, lactose, urea and many other novel traits. A new indicator trait for ketosis, KetoMIR, which is based on clinical ketosis diagnoses and MIR-predicted traits, was developed by the Regional State Association for Performance and Quality Inspection in Animal Breeding of Baden Württemberg in 2015. The KetoMIR result is available for each cow at milk recording during the first 120 days in milk and presented to farmers in three classes: 1 = low ketosis risk, 2 = moderate ketosis risk and 3 = high ketosis risk. The aim of the current study was to analyze the phenotypic relationships between KetoMIR and milk yield, fertility and health at the herd level. Annual herd reports from 12,909 herds with an average herd size of 27 cows were available for the analyses. Overall, the mean incidence of ketosis (KetoMIR risk class 2 or 3) at the herd level was 14.0%. Farms with the lowest ketosis risk (≤10% of cows in the herd with a moderate or high ketosis risk) differed in all variables from the farms with the highest ketosis risk (>50% of cows in the herd with a moderate or high ketosis risk). The increased ketosis risk based on KetoMIR was associated with lower average herd milk yield (-1,975 kg milk). Mean herd somatic cell count in first and higher lactations was increased by 60,500 and 134,400 cells/ml, respectively. The interval from calving to first service was prolonged by +36.5 days, as was the calving interval with +58.2 days. The newly developed KetoMIR trait may be used in ketosis prevention programs.

Genetic improvement of udder health in dairy cows is of high relevance as mastitis is one of the most prevalent diseases. Since it is known that the heritability of mastitis is low and direct data on mastitis cases are often not available in large numbers, auxiliary traits, such as somatic cell count (SCC) are used for the genetic evaluation of udder health. In previous studies, models to predict clinical mastitis based on mid-infrared (MIR) spectral data and a somatic cell count-derived score (SCS) were developed. Those models can provide a probability of mastitis for each cow at every test-day, which is potentially useful as an additional auxiliary trait for the genetic evaluation of udder health. Furthermore, MIR spectral data were used to estimate contents of lactoferrin, a glycoprotein positively associated with immune response. The present study aimed to estimate heritabilities (h2) and genetic correlations (ra) for clinical mastitis diagnosis (CM), SCS, MIR-predicted mastitis probability (MIRprob), MIR + SCS-predicted mastitis probability (MIRSCSprob) and lactoferrin estimates (LF). Data for this study were collected within the routine milk recording and health monitoring system of Austria from 2014 to 2021 and included records of approximately 54,000 Fleckvieh cows. Analyses were performed in two datasets, including test-day records from 5 to 150 or 5 to 305 days in milk. Prediction models were applied to obtain MIR- and SCS-based phenotypes (MIRprob, MIRSCSprob, LF). To estimate heritabilities and genetic correlations bivariate linear animal models were applied for all traits. A lactation model was used for CM, defined as a binary trait, and a test-day model for all other continuous traits. In addition to the random animal genetic effect, the fixed effects year-season of calving and parity-age at calving and the random permanent environmental effect were considered in all models. For CM the random herd-year effect, for continuous traits the random herd-test day effect and the covariate days in milk (linear and quadratic) were additionally fitted. The obtained genetic parameters were similar in both datasets. The heritability found for CM was expectedly low (h2 = 0.02). For SCS and MIRSCSprob, heritability estimates ranged from 0.23 to 0.25, and for MIRprob and LF from 0.15 to 0.17. CM was highly correlated with SCS and MIRSCSprob (ra = 0.85 to 0.88). Genetic correlations of CM were moderate with MIRprob (ra = 0.26 and 0.37) during 150 and 305 days in milk, respectively and low with LF (h2 = 0.10 and 0.11). However, basic selection index calculations indicate that the added value of the new MIR-predicted phenotypes is limited for genetic evaluation of udder health.

Currently, prevalence and incidence of claw lesions are used as parameters for benchmarking claw health. The aims of this study were to create a benchmarking system for claw health utilizing the claw health indicators Farm-Claw-Score (FCS) for the herd and Cow-Claw-Score (CCS) for the individual animal, and to benchmark claw health of the three predominant dairy cattle breeds in Austria. Claw health data from 17,642 cows from 508 Austrian dairy farms were analyzed. The CCS and FCS were calculated based on recorded claw lesions and their three severity levels using geometrically weighted scoring. The FCS of each of the dairy farms was classified into five percentile thresholds (P10, P25, P50, P75, P90), with the FCS calculated using the median value of CCS in each herd. Furthermore, claw health was benchmarked for three breeds (Fleckvieh, Holstein, Brown Swiss cows), using claw lesion prevalences and CCS values.When the median FCS was calculated, dairy farms in P50 and below had an FCS of 20.0, indicating very good claw health. However, P90 farms showed an FCS-MEDIAN of 67.5. Evaluation of the prevalences of the 14 claw lesions considered and the CCS values revealed that Fleckvieh cows (CCS-MEDIAN: 24.0), followed closely by Holstein cows (CCS-MEDIAN: 22.7) had significantly poorer claw health (P < 0.0001) compared to Brown Swiss cows (CCS-MEDIAN: 12.0). The use of CCS and FCS as primary claw health indicators allowed for a quick assessment of the current state of an individual cow and a dairy herd in a benchmarking system. Detailed information on the claw health of each animal and the dairy herd can be easily reviewed by examining diagnosis lists that display prevalences, particularly those related to lameness, in the respective electronic documentation systems.

We recently reported the ubiquitous occurrence of mycotoxins and their secondary metabolites in dairy rations and a substantial variation in the feeding management among Austrian dairy farms. The present study aimed to characterize to which extent these factors contribute to the fertility, udder health traits, and performance of dairy herds. During 2019 and 2020, we surveyed 100 dairy farms, visiting each farm 2 times and collecting data and feed samples. Data collection involved information on the main feed ingredients, nutrient composition, and the levels of mycotoxin and other metabolites in the diet. The annual fertility and milk data of the herds were obtained from the national reporting agency. Calving interval was the target criterion for fertility performance, whereas the percentage of primiparous and multiparous in the herd with somatic cell counts (SCC) above 200,000 cells/mL for impaired udder health. For each criterion, herds were classified into 3 groups: High/Long, Mid and Low/Short, with the cut-off corresponding to the < 25th and > 75th percentiles and the rest of the data, respectively. Accordingly, for the calving interval, the cut-offs for the Long and Short groups were ≥400 d and ≤380 d, for the udder health in primiparous cows were ≥20% and ≤8% of the herd, and for the udder health in multiparous cows were ≥35% and ≤20% % of the herd, respectively. Quantitative approaches were further performed to define potential risk factors in the herds. The High SCC group had higher dietary concentrations of enniatins (ENNs) (2.8 vs. 1.62 mg/cow/day), deoxynivalenol (DON) (4.91 vs. 2.3 mg/cow/day), culmorin (CUL) (9.48 vs. 5.72 mg/cow/day), beauvericin (BEA) (0.32 vs. 0.18 mg/cow/d) and siccanol (13.3 vs. 5.15 mg/cow/d), total Fusarium metabolites (42.8 vs. 23.2 mg/cow/d) and used more corn silage in the ration (26.9 vs. 17.3% diet DM) compared with the Low counterparts. BEA was the most substantial contributing variable among the Fusarium metabolites, as indicated by logistic regression and modeling analyses. Logistic analysis indicated that herds with high proportions of cows with milk fat-to-protein ratio >1.5 had an increased odds for a longer calving interval, which was found significant for primiparous cows (odds ratio = 5.5, 95% confidence interval (CI) = 1.65 - 21.7). As well, herds with high proportions of multiparous cows showing levels of milk urea nitrogen > 30 mg/dL had an increased odds for longer calving intervals (odds ratio = 2.96, CI = 1.22 - 7.87). In conclusion, the present findings suggest that dietary contamination of Fusarium mycotoxins (especially emerging ones), likely due to increased use of corn silage in the diet, seems to be a risk factor for impairing the udder health of primiparous cows. Mismatching dietary energy and protein supply of multiparous cows contributed to reduced herd fertility performance.

Genetic improvement of farm animals, especially selection within breeds focussed on high production and efficiency, is often cited as a potential threat to animal welfare. However, many animal welfare issues can be addressed, at least partially, by animal breeding and genetics. In this chapter, we explore the relationship between genetic selection and animal welfare, the strategies and tools for genetic improvement and how they can contribute to improved animal welfare. A growing public awareness of animal welfare and environmental issues has led to breeding goals being broadened beyond farmer profitability. As animal welfare and behaviour are complex and multi-factorial, so the emergence of selection indices that include a large number of traits to optimise animal welfare in a way that is consistent with enterprise sustainability for the farmer is necessary. This trend is likely to continue and will be aided by the advent of new technologies for measuring animal welfare in conjunction with DNA-based predictions of genetic merit (genomic selection). The dairy cattle industry has been exemplary for the application of genomic selection, in addition to enabling selection decisions to be made earlier in life, it can be used to select for traits where it was not possible to select for previously. These include important welfare-related traits, such as improved disease resistance and heat tolerance. Dairy cattle breeding is a very international activity with just a few breeding companies dominating the market in semen for the most numerous breeds, especially the Holstein. Consequently, genetic diversity within breeds is diminishing and although genetic gain has been significant, the rate of inbreeding now presents itself as a threat to the future success of breeding programmes. A greater emphasis on diversity in breeding programmes and the traits under selection is needed as major themes in research and application. Innovation in methods to measure these new traits, (e.g. molecular phenotyping, sensor development, digitalisation data science, etc.) could dramatically transform selection for animal welfare, as these technologies can enable large-scale objective measurements of animal behaviours. In addition to animal-based outcome measures, factors like housing, feeding, specific management practices pose other risks to welfare. Risk factors and their interactions have an impact on the development of diseases or other challenges to welfare. Collaborative efforts between animal behaviour scientists, geneticists, engineers, data scientists, and others will potentially provide solutions to these challenges.

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.

This study aimed to develop a tool to detect mildly lame cows by combining already existing data from sensors, AMSs, and routinely recorded animal and farm data. For this purpose, ten dairy farms were visited every 30–42 days from January 2020 to May 2021. Locomotion scores (LCS, from one for nonlame to five for severely lame) and body condition scores (BCS) were assessed at each visit, resulting in a total of 594 recorded animals. A questionnaire about farm management and husbandry was completed for the inclusion of potential risk factors. A lameness incidence risk (LCS ≥ 2) was calculated and varied widely between farms with a range from 27.07 to 65.52%. Moreover, the impact of lameness on the derived sensor parameters was inspected and showed no significant impact of lameness on total rumination time. Behavioral patterns for eating, low activity, and medium activity differed significantly in lame cows compared to nonlame cows. Finally, random forest models for lameness detection were fit by including different combinations of influencing variables. The results of these models were compared according to accuracy, sensitivity, and specificity. The best performing model achieved an accuracy of 0.75 with a sensitivity of 0.72 and specificity of 0.78. These approaches with routinely available data and sensor data can deliver promising results for early lameness detection in dairy cattle. While experimental automated lameness detection systems have achieved improved predictive results, the benefit of this presented approach is that it uses results from existing, routinely recorded, and therefore widely available data.

The aim of this study was to investigate the prevalence of ESBL/AmpC-producing E. coli and the resistance pattern of commensal E. coli, as well as the link between the use of antibiotics (AMU) and the occurrence of resistance in E. coli on Austrian dairy farms. AMU data from 51 farms were collected over a one-year period in 2020. Fecal samples were collected from cows, pre-weaned and weaned calves in 2020 and 2022. Samples were then analyzed using non-selective and selective agar plates, E. coli isolates were confirmed by MALDI-TOF analysis. Broth microdilution was used for antimicrobial susceptibility testing. The AMU of each farm was quantified as the number of Defined Daily Doses (nDDDvet) and Defined Course Doses (nDCDvet) per cow and year. Cephalosporins (mean 1.049; median 0.732 DDDvet/cow/year) and penicillins (mean 0.667; median 0.383 DDDvet/cow/year) were the most frequently used antibiotics on these farms, followed by tetracyclines (mean 0.275; median 0.084 DDDvet/cow/year). In 2020, 26.8% of the E. coli isolated were resistant to at least one antibiotic class and 17.7% of the isolates were classified as multidrug resistant (≥3 antibiotic classes). Out of 198 E. coli isolates, 7.6% were identified as extended-spectrum/AmpC beta-lactamase (ESBL/AmpC) producing E. coli. In 2022, 33.7% of E. coli isolates showed resistance to at least one antibiotic and 20.0% of isolates displayed multidrug resistance. Furthermore, 29.5% of the samples carried ESBL/AmpC-producing E. coli. In 2020 and 2022, the most frequently determined antibiotic resistances among commensal E. coli isolates were to tetracyclines, sulfonamides and penicillins. In addition, pre-weaned calves had the highest resistance rates in both years. Statistical analyses showed a significant association between low and high use AMU classifications for penicillins (in nDDDvet/cow/year) and their respective resistance among commensal E. coli isolates in 2020 (p = 0.044), as well as for sulfonamide/trimethoprim (p = 0.010) and tetracyclines (p = 0.042). A trend was also noted between the total amount of antibiotics used on farm in 2020 (by nDDDvet/cow/year) and multidrug resistances in commensal E. coli isolated on farm that year (p = 0.067). In conclusion, the relationship between AMU and antimicrobial resistance (AMR) on dairy farms continues to be complex and difficult to quantify.

Bovine mastitis is the most commonly diagnosed disease of dairy cows worldwide and causes extensive economic losses to milk producers. Intramammary infection status before dry-off plays a decisive role with respect to udder health and milk yield in the subsequent lactation. The aim of this study was to compare the effect of antibiotic dry cow therapy (DCT) versus no treatment at dry-off on milk yield, somatic cell count (SCC), inflammation of the mammary gland (IMG), and the incidence of clinical mastitis in the subsequent lactation. Dairy herd data from 251 Austrian dairy farms were recorded over an observation period of 12 mo and subsequently analyzed. The data set included 5,018 dairy cows: 2,078 were treated with antibiotics (abDCT group) and 2,940 were not treated (noDCT group) at dry-off. The abDCT group was subdivided, based on the antimicrobial active substances used for drying off, into 4 different groups (penicillins, cloxacillin, cephalosporins, and rifaximin). Based on bacteriological culture results, infections were grouped into those caused by major, minor, and other pathogens. Additionally, the IMG was defined via SCC from milk recording data using a cutoff of 200,000 cells/mL before drying off and after calving. The incidence of clinical mastitis cases within 30 and 90 d in milk was calculated using veterinary diagnosis data. To investigate the effect of different dry cow therapies on the following parameters: milk yield, SCC, and diagnosed clinical mastitis cases, different linear mixed models were constructed. Overall, the abDCT group was determined to have a significantly higher milk yield over 305 d in milk in the subsequent lactation (increase of 6.18%), compared with the noDCT group (increase of 4.29%). Both groups (abDCT and noDCT) demonstrated a decrease in the first SCC after calving compared with the SCC before dry-off, although the treated cows had a significantly higher reduction. Regarding the different antibiotic groups, with exception of the rifaximin treated cows, all antibiotic groups showed a significant difference from not treated cows with respect to SCC. Additionally, we were able to demonstrate that cows with IMG before dry-off had a 2.073 times higher chance of an increased SCC (>200,000 cells/mL) after calving. With respect to the veterinary diagnosis data, neither the IMG before drying off nor the type of DCT had a significant influence on the probability of developing clinical mastitis within 30 or 90 d in milk. Only a small number of treatments was accompanied with a bacteriological examination before drying off. However, the existing data in this study indicates that the intramammary infection status before dry-off in combination with different dry cow treatments influences udder health and milk yield after calving. Nevertheless, further studies with larger data sets of bacteriological examinations are necessary to enable a more in-depth investigation into the effects of different antibiotic substances used for DCT.

Monitoring for mastitis on dairy farms is of particular importance, as it is one of the most prevalent bovine diseases. A commonly used indicator for mastitis monitoring is somatic cell count. A supplementary tool to predict mastitis risk may be mid-infrared (MIR) spectroscopy of milk. Because bovine health status can affect milk composition, this technique is already routinely used to determine standard milk components. The aim of the present study was to compare the performance of models to predict clinical mastitis based on MIR spectral data and/or somatic cell count score (SCS), and to explore differences of prediction accuracies for acute and chronic clinical mastitis diagnoses. Test-day data of the routine Austrian milk recording system and diagnosis data of its health monitoring, from 59,002 cows of the breeds Fleckvieh (dual purpose Simmental), Holstein Friesian and Brown Swiss, were used. Test-day records within 21 days before and 21 days after a mastitis diagnosis were defined as mastitis cases. Three different models (MIR, SCS, MIR + SCS) were compared, applying Partial Least Squares Discriminant Analysis. Results of external validation in the overall time window (−/+21 days) showed area under receiver operating characteristic curves (AUC) of 0.70 when based only on MIR, 0.72 when based only on SCS, and 0.76 when based on both. Considering as mastitis cases only the test-day records within 7 days after mastitis diagnosis, the corresponding areas under the curve were 0.77, 0.83 and 0.85. Hence, the model combining MIR spectral data and SCS was performing best. Mastitis probabilities derived from the prediction models are potentially valuable for routine mastitis monitoring for farmers, as well as for the genetic evaluation of the trait udder health.