Automated measurements of the ratio of concentrations of methane and carbon dioxide, [CH₄]:[CO₂], in breath from individual animals (the so-called "sniffer technique") and estimated CO₂ production can be used to estimate CH₄ production, provided that CO₂ production can be reliably calculated. This would allow CH₄ production from individual cows to be estimated in large cohorts of cows, whereby ranking of cows according to their CH₄ production might become possible and their values could be used for breeding of low CH₄-emitting animals. Estimates of CO₂ production are typically based on predictions of heat production, which can be calculated from body weight (BW), energy-corrected milk yield, and days of pregnancy. The objectives of the present study were to develop predictions of CO₂ production directly from milk production, dietary, and animal variables, and furthermore to develop different models to be used for different scenarios, depending on available data. An international dataset with 2,244 records from individual lactating cows including CO₂ production and associated traits, as dry matter intake (DMI), diet composition, BW, milk production and composition, days in milk, and days pregnant, was compiled to constitute the training dataset. Research location and experiment nested within research location were included as random intercepts. The method of CO₂ production measurement (respiration chamber [RC] or GreenFeed [GF]) was confounded with research location, and therefore excluded from the model. In total, 3 models were developed based on the current training dataset: model 1 ("best model"), where all significant traits were included; model 2 ("on-farm model"), where DMI was excluded; and model 3 ("reduced on-farm model"), where both DMI and BW were excluded. Evaluation on test dat sets with either RC data (n = 103), GF data without additives (n = 478), or GF data only including observations where nitrate, 3-nitrooxypropanol (3-NOP), or a combination of nitrate and 3-NOP were fed to the cows (GF+: n = 295), showed good precision of the 3 models, illustrated by low slope bias both in absolute values (-0.22 to 0.097) and in percentage (0.049 to 4.89) of mean square error (MSE). However, the mean bias (MB) indicated systematic overprediction and underprediction of CO₂ production when the models were evaluated on the GF and the RC test datasets, respectively. To address this bias, the 3 models were evaluated on a modified test dataset, where the CO₂ production (g/d) was adjusted by subtracting (where measurements were obtained by RC) or adding absolute MB (where measurements were obtained by GF) from evaluation of the specific model on RC, GF, and GF+ test datasets. With this modification, the absolute values of MB and MB as percentage of MSE became negligible. In conclusion, the 3 models were precise in predicting CO₂ production from lactating dairy cows.

The aim of this study was to develop a basic model to predict enteric methane emission from dairy cows and to update operational calculations for the national inventory in Norway. Development of basic models utilized information that is available only from feeding experiments. Basic models were developed using a database with 63 treatment means from 19 studies and were evaluated against an external database (n = 36, from 10 studies) along with other extant models. In total, the basic model database included 99 treatment means from 29 studies with records for enteric CH₄ production (MJ/day), dry matter intake (DMI) and dietary nutrient composition. When evaluated by low root mean square prediction errors and high concordance correlation coefficients, the developed basic models that included DMI, dietary concentrations of fatty acids and neutral detergent fiber performed slightly better in predicting CH₄ emissions than extant models. In order to propose country-specific values for the CH₄ conversion factor Y_m (% of gross energy intake partitioned into CH₄) and thus to be able to carry out the national inventory for Norway, the existing operational model was updated for the prediction of Y_m over a wide range of feeding situations. A simulated operational database containing CH₄ production (predicted by the basic model), feed intake and composition, Y_m and gross energy intake (GEI), in addition to the predictor variables energy corrected milk yield and dietary concentrate share were used to develop an operational model. Input values of Y_m were updated based on the results from the basic models. The predicted Y_m ranged from 6.22 to 6.72%. In conclusion, the prediction accuracy of CH₄ production from dairy cows was improved with the help of newly published data, which enabled an update of the operational model for calculating the national inventory of CH₄ in Norway.

Large language models (LLM) hold significant promise to transform dairy science by enhancing research interpretation, supporting decision making, and improving knowledge dissemination. However, without proper systematic design, LLM may generate irrelevant or factually inaccurate responses for domain-specific questions. Moreover, most existing LLM and related tools are not tailored to the needs of the dairy domain, limiting their practical application within the field. To demonstrate the feasibility and practical value of embracing LLM in dairy science, we developed a 2-component agentic system: (1) a decision-support chatbot grounded in the Journal of Dairy Science (JDS) for science-backed insights and (2) a natural language interface for interacting with academic models and visualizing prediction results. All publicly available JDS abstracts and associated metadata dating back to 1917 were compiled using the PubMed application programming interface, forming a scientific knowledge base that enables the chatbot to answer user questions. A retrieval-augmented generation framework was implemented to ensure that responses generated by LLaMA (a LLM developed by Meta) were well-grounded in peer-reviewed literature, with the 5 most relevant sources cited alongside each answer. To address questions beyond the coverage of JDS literature, a web search agent was incorporated into the system to retrieve supplementary information from external online sources. Grading agents, powered by DBRX (a LLM developed by Databricks), were incorporated to evaluate the credibility and relevance of LLM-generated content to mitigate the risk of misinformation or hallucinated responses. The second component of the system facilitates natural language interaction with MilkBot, a published Bayesian milk yield prediction model. After users submit questions in plain language, the system converts the question into model parameters for MilkBot, executes the model prediction, and uses the predicted output to generate visualizations. This work demonstrates the capability of LLM to serve as intuitive, user-friendly interfaces for dairy-specific models. To our knowledge, this is the first chatbot prototype that integrates large-scale information from scientific literature, web-based resources, and academic models, along with self-evaluation capability, to provide dairy-specific insights to scholars, consultants, and farmers. However, challenges remain to realize the full value of LLM-assisted decision making, such as the lack of region-specific data to tailor the answers to the local circumstances, the need for more robust measures to protect data security and privacy, and the need to integrate additional functions to enable more comprehensive decision support.

This study aimed to quantify the effects of dietary inclusion of tannin-rich pomegranate peel (PP) on intake, methane and nitrogen (N) losses, and metabolic and health indicators in dairy cows. Four multiparous, late-lactating Brown Swiss dairy cows (796 kg body weight; 29 kg/d of energy corrected milk yield) were randomly allocated to 3 treatments in a randomized cyclic change-over design with 3 periods, each comprising 14 d of adaptation, 7 d of milk, urine, and feces collection, and 2 d of methane measurements. Treatments were formulated using PP that replaced on a dry matter (DM) basis 0% (control), 5%, and 10% of the basal mixed ration (BMR) consisting of corn and grass silage, alfalfa, and concentrate. Gaseous exchange of the cows was determined in open-circuit respiration chambers. Blood samples were collected on d 15 of each period. Individual feed intake as well as feces and urine excretion were quantified, and representative samples were collected for analyses of nutrients and phenol composition. Milk was analyzed for concentrations of fat, protein, lactose, milk urea N, and fatty acids. Total phenols and antioxidant capacity in milk and plasma were determined. In serum, the concentrations of urea and bilirubin as well as the activities of alanine aminotransferase (ALT), aspartate aminotransferase, glutamate dehydrogenase, alkaline phosphatase, and γ-glutamyl transferase were measured. The data was subjected to ANOVA with the Mixed procedure of SAS, with treatment and period as fixed and animal as random effects. The PP and BMR contained 218 and 3.5 g total extractable tannins per kg DM, respectively, and thereof 203 and 3.3 g hydrolyzable tannins. Total DM intake, energy corrected milk, and methane emission (total, yield, and intensity) were not affected by PP supplementation. The proportions of C18:2 n-6 and C18:3 n-3 in milk increased linearly as the amount of PP increased in the diet. Milk urea N, blood urea N, and urinary N excretion decreased linearly with the increase in dietary PP content. Total phenols and antioxidant capacity in milk and plasma were not affected by the inclusion of PP. The activity of ALT increased in a linear manner with the inclusion of PP. In conclusion, replacing up to 10% of BMR with PP improved milk fatty acid composition and alleviated metabolic and environmental N load. However, the elevated serum ALT activity indicates an onset of liver stress even at 5% PP, requiring the development of adaptation protocols for safe inclusion of PP in ruminant diets.

Summary: The dairy sector is operating in an environment of rapid technological change, driven by advances in big data and artificial intelligence (AI). To remain competitive and sustainable, farms should adopt intelligent, scalable, and privacy-preserving solutions. This paper presents a comprehensive ecosystem that combines multimodal AI, edge computing, and federated learning to enable real-time, data-driven decision making on farms. With emphasis on privacy, local relevance, and research integration, the framework would enable autonomous AI agents to support everyday farm decisions and long-term planning. The graphical abstract depicts a schematic representation of the proposed framework for a future dairy data ecosystem, illustrating interactions among commercial and research farms, edge devices, and the cloud intelligence layer (enterprise and research and development [R&D] clouds). It shows data and model flows, with solid arrows pointing away from farms and dashed arrows pointing toward farms. E = edge layer; FL = federated learning; RAW = raw data.

This study aimed to assess relationships of enteric methane (CH₄) yield (g/kg of DMI) with immune response, feed efficiency (ECM/DMI), and rumen microbiome in dairy cows, both in early and in late lactation. The DMI, BW, ECM yield, and CH₄ emission were measured in respiration chambers in early (n = 20, 32 ± 7 DIM) and nonpregnant late lactating (n = 14, 359 ± 90 DIM) multiparous Holstein cows. The in vitro immune response was studied in response to (1) LPS using whole blood, and (2) phytohemagglutinin and concanavalin A using peripheral blood mononuclear cells. The DNA was extracted from rumen content samples (esophageal tubing) for 16S rRNA microbial analysis. Cows were divided retrospectively into an equal number of cows with low (LMY) and high (HMY) CH₄ yield within each lactation group. In early lactation, CH₄ yields in LMY (n = 10) and HMY cows (n = 10) were on average (±SD) 18.8 ± 1.4 and 23.5 ± 2.8 g/kg of DMI, respectively. In late lactation, CH₄ yields in LMY (n = 7) and HMY cows (n = 7) were 20.8 ± 2.0 and 23.6 ± 1.7 g/kg of DMI, respectively. Statistical analysis was performed separately for each lactation group. In early lactation, we found that whole blood and isolated peripheral blood mononuclear cells from LMY compared with HMY animals were less responsive to stimulants in vitro. In addition, feed conversion efficiency was lower in LMY than HMY cows, and the relative abundance of the archaeal genus Methanospaera and the bacterial genus Marvinbryantia were higher. In late lactation, we observed no differences in immune response and feed conversion efficiency between LMY and HMY cows. Still, in LMY cows several bacterial genera including Prevotella 7, Ruminococcus gauvreauii group, and Shuttleworthia were enriched, whereas in HMY cows Methanobrevibacter, Veillonellaceae UCG-001, Succinivibrionaceae UCG-002, Rikenellaceae RC9, and CAG-352 were enriched. The results indicate that in early lactation the animals with low CH₄ yield reach energy balance faster, at the expense of an inadequate immune response. Meanwhile, increased CH₄ yield in early lactation may reflect higher rumen fermentation activity, fostering feed efficiency and energy availability for supporting immune function.