Journal of Dairy Science最新文献

{"title":"Mesoscale fractal whey protein particles derived from microscale linear-shaped protein assemblies (Part 2): Foaming properties and heat stability.","authors":"Hanyu Shi, Haotian Zheng","doi":"10.3168/jds.2025-27190","DOIUrl":"https://doi.org/10.3168/jds.2025-27190","url":null,"abstract":"<p><p>This study investigates the functionality of mesoscale whey protein particles (WPP) derived from fiber- and ribbon-shaped whey protein assemblies produced via a liquid antisolvent precipitation-based method. The air-water (A-W) interfacial characteristics, foaming properties, and heat stability of WPP were evaluated and compared with the original whey protein source, whey protein isolate (WPI). Adsorption dynamics and dilatational rheology at the A-W interface were characterized using pendant drop and oscillating drop methods, respectively. Foamability and foam stability were assessed using a dynamic foam analyzer, and heat stability was evaluated by examining changes in particle size distribution (PSD) profiles and turbidity before and after heat treatment at 95°C for 5 min. Whey protein particles achieved a quasi-equilibrium surface pressure comparable to WPI after 3 h of adsorption but showed lower dilatational elastic moduli during dilatational deformation. Although WPI exhibited faster surface adsorption, no significant difference was observed in the rate constant of penetration (k_p) between WPP and WPI. Compared with WPI, reconstituted freeze-dried WPP dispersions did not improve foamability, likely due to slower A-W interface adsorption associated with their larger particle sizes; however, they demonstrated enhanced foam stability, evidenced by a longer 75%-volume lifetime. This improvement is likely attributed to the entrapment of WPP within the lamellae and Plateau borders of foam structure, which may increase local viscosity and block the liquid drainage channels, thereby retarding foam collapse. Moreover, WPP exhibited minimal changes in their PSD and turbidity after heat treatment, suggesting enhanced heat stability relative to WPI.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145273460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Graduate Student Literature Review: Limitations in feeding red seaweed Asparagopsis species for enteric methane mitigation in ruminants.","authors":"Ibrahim Ahmad, Richard P Rawnsley, John P Bowman, Apeh A Omede","doi":"10.3168/jds.2025-26915","DOIUrl":"https://doi.org/10.3168/jds.2025-26915","url":null,"abstract":"<p><p>Enteric methane (CH₄) emission from ruminant livestock results in a loss of digestible feed energy, which has significant implications for animal production and contributes to global warming. Digestible feed energy losses and GHG emissions are sufficient justifications for devising strategies that reduce enteric CH₄ emissions from ruminants. Of the available and horizon GHG mitigation strategies for ruminants, there is currently considerable enthusiasm in the use of feed additive methanogenesis inhibitors. Asparagopsis, as a feed additive, is very effective at mitigating enteric CH₄ emissions from ruminants. Although research has focused on the potential of this feed additive to mitigate enteric CH₄, Asparagopsis supplementation has frequently been accompanied by a noticeable decline in DMI. This review summarizes the current literature on challenges and implications that persist in Asparagopsis feeding trials in ruminants. These include constraints in the experimental design and effects of the novel ingredient on dihydrogen (H₂) metabolism, which could affect rumen functions and, consequently, DMI. Concerns have been also raised regarding the cost, potential side effects on animal health, and the safety of Asparagopsis feed additives. Antimethanogenic activity must be reconciled with other nutritional considerations, including a reduction in rumen fermentation efficiency, which could contribute to a decrease in feed intake and indirectly affect animal health and productivity. Therefore, this synthesis provides essential information for the livestock industry to develop standardized protocols that could minimize the negative effects of dietary Asparagopsis on feed intake and productivity.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145273278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The effect of bovine leukemia virus infection on health and growth of nonreplacement dairy calves.","authors":"K Broadfoot, F van der Meer, R Couto Serrenho, F Pharo, A Keunen, D L Renaud","doi":"10.3168/jds.2025-27122","DOIUrl":"https://doi.org/10.3168/jds.2025-27122","url":null,"abstract":"<p><p>The objective of this cohort study was to determine if bovine leukemia virus (BLV) positivity, identified using quantitative PCR (qPCR), was associated with health and growth outcomes in nonreplacement dairy calves. A commercial calf-raising facility located in Southwestern Ontario was used, where a total of 768 male dairy calves (726 Holstein, 42 crossbred), estimated to be 3 to 10 d of age, had 3 blood samples taken at 1 d, 30, and 84 after arrival to the facility. Whole blood EDTA samples were taken and sent to the University of Calgary Faculty of Veterinary Medicine for qPCR analysis to determine positivity for BLV. Calves were determined to be positive using a baseline threshold of >200 relative fluorescence unit by <40 cycles. All calves were health scored 2 times daily for fecal consistency and respiratory disease, using the UC Davis respiratory scoring chart. Additionally, calves were weighed at time of arrival and weekly thereafter until they left the facility at 84 d. Mixed linear regression models were built to assess ADG (kg/d) over 84 d and feed efficiency (ME/kg of gain), whereas a Poisson model was built to assess the number of observations with a respiratory score of ≥5. Further, a negative binomial model was built to assess the number of observations with diarrhea (score of ≥2), and a zero-truncated Poisson model was built to assess the number of observations with no clinical signs of respiratory disease (e.g., respiratory score of 0). Treatment for respiratory disease and diarrhea was also analyzed using a Cox proportional hazard model. Two calves had incomplete blood sampling and were therefore removed from the analysis, leaving a total of 766 calves. A total of 43 (5.9%) calves tested positive for BLV via qPCR testing over the 84-d period. Calves that were BLV-positive were associated with a greater number of observations with a respiratory score of ≥5 (incidence rate ratio [IRR] = 1.91, 95% CI: 1.69 to 2.17) and a lower number of observations with a score of 0 (IRR = 0.98, 95% CI: 0.95 to 1.00). There was no difference identified between BLV positivity and ADG (kg/d), feed efficiency (ME/kg of gain), or diarrhea. Overall, this study identifies an association between BLV status and respiratory disease occurrence in preweaning calves. Although the underlying mechanisms and long-term effects remain unclear, these findings highlight the need for further research to determine whether BLV plays a causal role in respiratory disease in calves.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145273443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"β-Hydroxybutyrate inhibits the spontaneous apoptosis of neutrophils through a degranulation-macrophage-1 antigen axis in cows with subclinical ketosis.","authors":"Kexin Wang, Guowen Liu, Ke Zhang, Zhiyong Li, Yan Ren, Shang Jiang, Xiliang Du, Lin Lei, Wenwen Gao, Zhe Wang, Xinwei Li, Yuxiang Song","doi":"10.3168/jds.2025-26516","DOIUrl":"10.3168/jds.2025-26516","url":null,"abstract":"<p><p>Subclinical ketosis (SCK), a common metabolic disorder in dairy cows during the peripartum period, is accompanied by systemic inflammation and elevated circulating BHB. β-Hydroxybutyrate contributes to the development of systemic inflammation by inhibiting the spontaneous apoptosis of neutrophils in SCK cows. However, the underlying molecular mechanism is still unclear. This study investigated the role of the degranulation-macrophage-1 antigen (Mac-1) axis in this process. The results demonstrated that BHB promoted the degranulation process and activated membrane Mac-1 (CD11b/CD18) on neutrophils both ex vivo and in vitro. Neutrophils were isolated from healthy cows and treated with 2 mM BHB, a specific degranulation inhibitor, or a Mac-1 blocking antibody in vitro. We found that the released granule contents during degranulation activated the intracellular Mac-1 signaling pathway, which was involved in BHB-mediated inhibition of bovine neutrophil apoptosis. Overall, our findings reveal that the degranulation-Mac-1 axis plays a critical role in regulating BHB-induced inhibition of spontaneous neutrophil apoptosis in dairy cows with SCK.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":"11403-11417"},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144768198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Genetic background of calving ease in beef-on-dairy.","authors":"Arielly O Garcia, Ashley A Mikush, John B Cole, Shogo Tsuruta, Ignacy Misztal, Simone E F Guimarães, Daniela Lourenco","doi":"10.3168/jds.2025-26503","DOIUrl":"https://doi.org/10.3168/jds.2025-26503","url":null,"abstract":"<p><p>A common practice in dairy herds is to breed females not selected as replacement heifers to beef bulls. This increases the market value of the surplus calves sold for beef purposes. Some beef breed associations have built selection indices focusing mainly on carcass traits; however, calving ease (CE) is also an important trait, given that crossbreeding with beef bulls can change gestation patterns (e.g., gestation length) or calf conformation (e.g., weight and size), generating a negative effect on the health, and consequently on the production, of the cows. We used linear and threshold animal models to estimate genetic parameters and breeding values for direct and maternal additive effects for CE in beef-on-dairy crosses, considering only the first or the first 3 lactations. We analyzed 231K CE records in the first lactation and 1.2 million in the first 3 lactations from Holstein and Jersey cows inseminated with Angus, Charolais, or Simmental semen. Although CE was scored in 5 categories, we reduced this to a binary trait (1 = easy and 2, 3, 4, 5 = difficult). The average incidence of difficult calving (scores ≥2) was ∼15%. Direct and maternal heritabilities for the linear (threshold) model were 0.01 ± 0.002 (0.01 ± 0.001) and 0.02 ± 0.002 (0.04 ± 0.004), respectively, using the first lactation, and equal to 0.01 ± 0.002 (0.01 ± 0.009) and 0.19 ± 0.002 (0.26 ± 0.006), respectively, considering the first 3 lactations. Maternal heritabilities were always greater than the direct ones. Maternal heritabilities were inflated when we considered more than one lactation, most likely because of a confounding with the maternal permanent environmental effect that could not be estimated. Linear and threshold models provided similar direct EBV rankings, with a correlation of at least 0.86 when considering all different breeds; for maternal effect, it was high for dairy breeds (>0.9) and close to zero in beef breeds. Validation metrics were better for the linear model with only first lactation records. Although with the small direct heritabilities, the results showed that direct genetic variability exists, and that it would be possible to select beef bulls based on their direct EBV for CE in beef-on-dairy systems. One of the challenges in beef-on-dairy analyses is the lack of pedigree depth on the sire side. When this is the case, we suggest using linear models considering only the first lactation to evaluate CE, given that EBV are highly correlated with those obtained by the threshold model but are less biased and converge almost 10 times faster, proving to be more efficient for routine genetic evaluations.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Characterization of a novel heat tolerance trait and subsequent haplotype block-based analysis to identify associated regions in Dutch Holstein cattle.","authors":"T Pook, M L van Pelt, J Vandenplas, I Adriaens, L Zetouni, C Orrett, Y de Haas, C Kamphuis, B Gredler-Grandl","doi":"10.3168/jds.2025-26870","DOIUrl":"https://doi.org/10.3168/jds.2025-26870","url":null,"abstract":"<p><p>Heat stress is a major environmental challenge affecting dairy cattle, leading to behavioral changes, production losses, and welfare concerns. As heat stress events intensify and become more frequent due to climate change, identifying animals that are able to maintain high production levels during heat stress events, commonly referred to as heat tolerance, is crucial for sustainable dairy production. In this study, we develop a pipeline to quantify the population-wise impact of heat stress on a dairy cattle population and subsequently define individual-based heat tolerance traits. Data from 677,318 Dutch Holstein cows were analyzed, including 15.6 million mid-infrared spectra and 762 million records from automated milking systems. An iterative approach using kernel regression was employed to estimate the population-wise effects of heat stress. Results indicate that fat and protein percentages decrease approximately linearly with increasing temperature-humidity index (THI), with an absolute reduction of 0.3% as THI increases from 30 to 70. In contrast, milk yield remains stable until a THI of 60, after which production losses increase quadratically, reaching a loss of 5.0% at a THI of 75. We subsequently define the heat tolerance phenotype of an animal as the slope from a linear regression model of the residuals of the population-wise models against THI for milk yield, concentration of fat, protein, lactose, and specific fatty acids. Compared with reaction-norm models, individual records per cow are combined into one joint record before model fitting, thus reducing computing times and allowing more flexibility in the design of the model. Heritabilities for heat tolerance traits ranged from 0.05 to 0.12, and genetic variances indicate substantial potential for breeding as an improvement of the population by 1 genetic standard deviation would already offset 69% of the losses in fat percentage, 65% in protein percentage, and 11% in milk yield. Heat tolerance based on milk yield showed favorable correlations with most commercial traits, whereas heat tolerance based on fat and protein percentage showed negative correlations with health and resilience. A GWAS using both SNPs and haplotype blocks from the software HaploBlocker identified potential QTL across the genome, with particularly strong signals on BTA5, 14, and 20. These findings support the long-term potential of genetic improvement through breeding for heat tolerance but highlight the need for complementary management strategies to mitigate heat stress impacts in the short term.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effects of calcium peroxide or biochar-enzyme feed additives on milk production, enteric methane emissions, and ruminal microbiota in Nordic Red dairy cows.","authors":"J Vattulainen, A R Bayat, T Stefański, M Rinne, I Tapio","doi":"10.3168/jds.2025-27123","DOIUrl":"https://doi.org/10.3168/jds.2025-27123","url":null,"abstract":"<p><p>Expanded basic research is needed to discover and develop wider selection of dietary additives that are economically feasible without compromising animal performance or health. The objective of this study was to evaluate the effect of 2 feed additives as methane-mitigating agents in milk production systems. Four multiparous Nordic Red dairy cows were assigned to a 4 × 4 Latin square experiment with four 28-d periods using respiration chambers. The control diet (CON) consisted of grass silage and dietary concentrates mixed at forage-to-concentrate ratio of 65:35 on DM basis. The 3 experimental treatments consisted of the CON diet supplemented with 0.2% of biochar with fibrolytic enzymes and live yeast additive (BFE) or with 0.75% or 1.5% CaO₂ on a DM basis (CaPe1 and CaPe2, respectively). Calcium peroxide (CaPe) was included in the concentrate pellet; the mixture of biochar, fibrolytic enzymes and live yeast was added to the diet during TMR preparation; and diets were fed as TMR 4 times daily. Feeding BFE had minor effect on the parameters evaluated in the experiment. Feeding CaPe resulted in linear reductions in DMI, OMI, CP, ether extract (EE), NDF, and gross energy (GE) intake compared with CON. Yields of milk, ECM, fat, protein, lactose and total solids decreased linearly, but milk composition and SCC were not affected. Apparent total-tract digestibilities of DM, OM, CP, EE, NDF, and GE decreased linearly, whereas excretion of Ca and P in feces increased linearly with increasing CaPe level. We found that CaPe1 tended to decrease the molar proportion of acetate and increased that of propionate, whereas butyrate increased linearly. Dietary CaPe inclusion decreased daily CH₄ production (g/d) linearly by 15.0%, but CH₄ yield (g/kg DM or OM intake) and intensity (g/kg milk or ECM) were not affected. Hydrogen production (g/d) and yield (g/kg DMI) decreased at CaPe1 but plateaued at CaPe2. Feeding CaO₂ increased richness of ciliate protozoa and influenced rumen bacteria and ciliate protozoa community structure. No such effect was observed on archaea or anaerobic fungi. The feed additives BFE and CaPe were not effective CH₄-mitigating agents under the conditions of the present experiment.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A dynamic yak heifer pose estimation model based on keypoints detection for complex environmental monitoring.","authors":"Jiazhou Li, Yuxiang Yang, Yao Yao, Huawei Zou, Xi Guo, Jianxin Xiao, Rui Hu, Shijing Cheng, Yipeng Wang, Yingqi Peng, Zhisheng Wang","doi":"10.3168/jds.2025-27099","DOIUrl":"https://doi.org/10.3168/jds.2025-27099","url":null,"abstract":"<p><p>Behavior is an important indicator of yak heifer welfare and health status, with key behavioral patterns reflecting critical conditions, including fattening, reproduction, and disease. Computer vision-based pose estimation has become an essential technology for livestock behavior monitoring. This study developed a yak heifer pose estimation model named Multistage Feature Attention PVT-based Dynamic Yak Heifer Pose Estimation Model (MFPVT-YakHeifer) based on improved Pyramid Vision Transformer version 2 (PVT v2) and keypoint detection modeling. The proposed and compared models were trained with a novel YakPoseData set encompassing yak heifer images collected in different poses and environmental conditions. The results showed that the model achieved performance metrics of 90.41% mean average precision at an intersection over union (IoU) threshold of 0.5, 64.37% mean average precision at 0.75 IoU, 64.11% mean average precision, and 91.95% mean average recall at 0.5 IoU, all of which are higher than those of the other 4 benchmark models. Finally, the MFPVT-YakHeifer model has been deployed on edge computing device for livestock farm applications. The future work will focus on animal dataset expansion, real-time video analysis implementation, and computational efficiency optimization for edge deployment.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Incorporating heterosis effects enhances genetic evaluations for milk production and functional traits in Chilean crossbred dairy cows.","authors":"Álvaro G Morales, Sabrina T Amorim, Carlos Lizana, Rubén G Pulido, Mark D Hanigan, Rebecca R Cockrum, Gota Morota","doi":"10.3168/jds.2025-26445","DOIUrl":"https://doi.org/10.3168/jds.2025-26445","url":null,"abstract":"<p><p>To accurately predict breeding values, genetic evaluations must define an appropriate model that considers the main factors influencing the traits of interest. The challenge increases in crossbreeding populations, such as those found in Chile, where multiple breeds, including those from different countries, are used. In particular, the country of origin is not currently incorporated into genetic models used in Chile. The objectives of this study were 2-fold: (1) to evaluate different genetic models of increasing complexity to determine if the inclusion of a specific strain classification, strain proportions in crossbred animals, and heterosis effects could enhance genetic evaluations compared with the current model used in Chile; and (2) to determine the presence of heterosis effects among specific dairy genetic strains when crossed with Chilean Friesian cows, focusing on annual milk, fat, and protein production; SCS; and calving interval. A dataset of 1,429,132 records from 586,624 cows that calved between 1998 and 2018 was provided by the Chilean Agricultural and Services Cooperative (COOPRINSEM; Osorno, Chile). Using pedigree information, the proportion of each cattle breed according to its different countries of origin (genetic strain) was calculated for each animal. Subsequently, strains concerning Chilean Friesian cows were selected with a wide range of proportions. The final dataset included 8 strains: Chilean Friesian, French Holstein-Friesian, US-Holstein, US-Jersey, French Montbéliarde, New Zealand Holstein-Friesian, Swedish Red and White, and British Friesian, representing 369,755 observations collected between 2009 and 2018. Production (milk, fat, and protein production per lactation) and functional traits (SCS and calving interval) were also provided by COOPRINSEM. Four different models were tested: the current model using only 5 breed categories based on physical appearance (M1), a more precise classification considering 8 dairy strains categories (M2), M2 plus the inclusion of a cross-classified effect of crossbreeding proportion with strain categories (M3), and M3 plus the cross-classified effect of heterosis level with strain categories (M4). Our results show that M4 was the best model for analyzing information on crossbred dairy cows from Chile. Our study also suggests that there are specific effects for some strains, as well as heterosis effects between several strains and the Chilean Friesian.</p>","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224573","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Screening and selection of a machine learning algorithm for development of a model to select cows for clinical examination using data from automated health monitoring technologies and other predictors of cow health.","authors":"M M Perez, E Cabrera, C Rial, Y You, Y Wang, K Weinberger, D V Nydam, J O Giordano","doi":"10.3168/jds.2025-26511","DOIUrl":"https://doi.org/10.3168/jds.2025-26511","url":null,"abstract":"&lt;p&gt;&lt;p&gt;The objective of this study was to create a framework for training and selecting machine learning algorithms (MLA) to classify cow health status daily using data from multiple automated health monitoring systems (AHMS), including wearable and nonwearable sensors, combined with nonsensor data of potential value for predicting cow health. The work presented in this manuscript is part of a series of studies aimed at identifying a single candidate algorithm that, upon extensive refinement and further development, could be deployed in a commercial dairy operation to identify cows potentially affected by health disorders for clinical examination. Data from AHMS and other cow features and performance data, including the clinical health status of cows, were collected in a prospective cohort study including Holstein cows (n = 1,252). Data from AHMS used for MLA training included rumination, eating, and physical activity measured in the neck (neck sensor), temperature and physical activity measured in the reticulorumen (bolus sensor), physical activity and resting measured in the leg (leg sensor), and milk yield, milk electrical conductivity, and milk components (parlor sensors). Other non-AHMS data used were temperature and humidity index, cow and calving event features, and current and previous lactation performance and management indicators. The dataset included 22,415 cow-day records with 49 features. The dataset was split into training and testing sets in an 80:20 ratio, resulting in 17,887 and 4,528 cow-day records, respectively. Data imputation and standardization were applied automatically or manually. A diverse set of nondeep learning (n = 26) MLA were trained and compared using the open-source automated ML (AutoML) tool Lazy Predict Classifier (LZP). Upon selection of the best-performing nondeep learning algorithms (i.e., XGBoost, AdaBoost, Nearest Centroid, and Bernoulli Naive Bayes) from the pool tested with LZP, classifier algorithms were compared with more complex deep learning algorithms (multilayer perceptron, recurrent neural networks, long short-term memory networks, and gated recurrent unit models) not included in LZP. All algorithms underwent training and evaluation before selection of a single best-performing algorithm, using several metrics of performance. Ensemble learning models, particularly XGBoost, achieved the best performance and balanced results with a sensitivity of 82.4% and a precision of 42.6% combined with a specificity of 86.4% and a negative predictive value of 97.6%. This model also had the highest F1-score (0.56) and area under the curve (84.4%). The XGBoost algorithm also demonstrated robustness in handling missing data. Our comprehensive approach to MLA screening and selection enabled informed decisions in selecting a suitable algorithm for identifying cows for clinical examination based on the daily prediction of health disorder occurrence. The combination of the AutoML tool LZP and manual refinement and test","PeriodicalId":354,"journal":{"name":"Journal of Dairy Science","volume":" ","pages":""},"PeriodicalIF":4.4,"publicationDate":"2025-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145224589","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}