Comparison of machine learning methods to predict udder health status based on somatic cell counts in dairy cows

Revealing microbial diversity in buffalo milk with high somatic cell counts: implications for mastitis diagnosis and treatment

Mastitis is one of the most serious diseases that threatens the health of dairy animals. The somatic cell count (SCC) in milk is widely used to monitor mastitis. This study aimed to reveal the diversity of microorganisms in buffalo milk with high somatic cell count (SCC ≥ 3 × 105 cells/mL, n = 30) and low somatic cell count (SCC ≤ 5 × 104 cells/mL, n = 10), and identify the dominant bacteria that cause mastitis in a local buffalo farm. We also investigated the potential method to treat bacterial mastitis. The V3-V4 region of 16 S rDNA was sequenced. Results showed that, compared to the milk with low SCC, the high SCC samples showed lower microbial diversity, but a high abundance of bacteria and operational taxonomic units (OTUs). By in vitro isolation and culture, Escherichia coli, Staphylococcus aureus, and Klebsiella pneumoniae were found to be the leading pathogens, which is consistent with the 16 S rDNA sequencing data. We further isolated 3 of the main pathogens and established a pathogen detection method based on ELISA. In addition, the antibacterial effects of 10 antimicrobials and 15 Chinese herbal extracts were also investigated. Results showed that the microbial has developed tolerance to several of the antimicrobials. While the water extracts of Chinese herbal medicine such as Galla Chinensis, Coptis chinensis Franch, Terminalia chebula Retz, and Sanguisorba officinalis L can effectively inhibit the growth of main pathogens. This study provides novel insight into the microbial diversity in buffalo milk and a reference for the prevention, diagnosis, and treatment of mastitis.

The determination of mastitis severity at 4-level using Milk physical properties: A deep learning approach via MLP and evaluation at different SCC thresholds

Current research aims to generate an alternative model to classical methods in the determination of subclinical mastitis at 4 levels (healthy, suspicious, subclinical, and clinical). For this purpose, multilayer perceptron (MLP) artificial neural networks (ANN) was developed as test model. 5 variables from the physical properties of milk somatic cell count (SCC), electrical conductivity (EC), pH, density, and temperature at fore milking (TFM) were included in the model in the classification of mastitis. Model performance was validated on test data (%25) and compared with the multinomial logistic regression (MNLR). MLP model has shown a satisfactory performance with an accuracy of 95.14% and - 141 of AIC score better than the control model (MNLR) of 80.27% and - 133 AIC despite using higher number of parameters (104). Since the main problem is to diagnose subclinical mastitis, which does not cause any visible symptoms, it was important to distinguish between absolute subclinical (suspicious excluded positives) and absolute healthy (suspicious included positives) ones. Therefore, optimum cut-off threshold was evaluated for these two different scenarios with only variable SCC the gold standard indicator of subclinical mastitis and results were compared in the interpretation of model performance. The results show that the 5-variable MLP model exhibits a high sensitivity of 93.22% (AUC = 0.95 for healthy ones) at low cutoff thresholds as well. New studies should provide a better understanding by evaluating economics, sustainability, animal welfare and health aspects together to determine the optimal threshold value.

Performance comparison of machine learning models used for predicting subclinical mastitis in dairy cows: Bagging, boosting, stacking, and super-learner ensembles versus single machine learning models

Mastitis has a substantial impact on the dairy industry across the world, causing dairy producers to suffer losses due to the reduced quality and quantity of produced milk. A further problem, related to this issue, is the excessive use of antibiotics that leads to the development of resistance in different bacterial strains. The growing consumer awareness oriented toward food safety and rational use of antibiotics has promoted the search for new methods of early identification of cows that may be at risk of developing the disease. Subclinical mastitis does not cause any visible changes to the udder or milk, and therefore it is more difficult to detect than clinical mastitis. The collection of large amounts of data related to milk performance of cows allows using machine learning (ML) methods to build models that could be used for classifying cows into healthy and at risk of subclinical mastitis. The data used for the purpose of this study included information from routine milk recording procedures. The dataset consisted of 19,856 records of 2,227 Polish Holstein-Friesian cows from 3 herds. The authors decided to use the approach of building ensemble ML models, in particular bagging, boosting, stacking, and super-learner models, and comparing them for accuracy of identification of disease-affected cows against single ML models based on the support vector machines, logistic regression, Gaussian Naive Bayes, k-nearest neighbors, and decision tree algorithms. The models were trained and evaluated based on the information recorded for herd 1 and using an 80:20 train-test split ratio according to animal ID (to avoid data leakage). The information recorded for herds 2 and 3 was only used to evaluate on unseen data models developed using the herd 1 dataset. Among the single ML models, the support vector machines model was found to be the most accurate in predicting subclinical mastitis at subsequent test day when used both for the training set (mean F1-score of 0.760) and the testing sets containing data for herds 1, 2, and 3 (F1-score of 0.778, 0.790, and 0.741 respectively). The gradient boosting model was found to be the best performing model among the ensemble ML models (F1-score of 0.762, 0.779, 0.791, and 0.723 for the training set and the testing sets, respectively). The super-learner model, featuring the most advanced design and logistic regression in the meta layer, achieved the highest mean F1-score of 0.775 during the cross validation; however, it was characterized by a slightly worse prediction accuracy of the testing sets (mean F1-score of 0.768, 0.790, and 0.693 for herds 1, 2 and 3 respectively). The study findings confirm the promising role of ensemble ML methods, which were found to be slightly superior with respect to most of the single ML models.

Comparison of Machine Learning Tree-Based Algorithms to Predict Future Paratuberculosis ELISA Results Using Repeat Milk Tests

Machine learning algorithms have been applied to various animal husbandry and veterinary-related problems; however, its use in Johne's disease diagnosis and control is still in its infancy. The following proof-of-concept study explores the application of tree-based (decision trees and random forest) algorithms to analyze repeat milk testing data from 1197 Canadian dairy cows and the algorithms' ability to predict future Johne's test results. The random forest models using milk component testing results alongside past Johne's results demonstrated a good predictive performance for a future Johne's ELISA result with a dichotomous outcome (positive vs. negative). The final random forest model yielded a kappa of 0.626, a roc AUC of 0.915, a sensitivity of 72%, and a specificity of 98%. The positive predictive and negative predictive values were 0.81 and 0.97, respectively. The decision tree models provided an interpretable alternative to the random forest algorithms with a slight decrease in model sensitivity. The results of this research suggest a promising avenue for future targeted Johne's testing schemes. Further research is needed to validate these techniques in real-world settings and explore their incorporation in prevention and control programs.

The use of machine learning to predict somatic cell count status in dairy cows post-calving

Udder health remains a priority for the global dairy industry to reduce pain, economic losses, and antibiotic usage. The dry period is a critical time for the prevention of new intra-mammary infections and it provides a point for curing existing intra-mammary infections. Given the wealth of udder health data commonly generated through routine milk recording and the importance of udder health to the productivity and longevity of individual cows, an opportunity exists to extract greater value from cow-level data to undertake risk-based decision-making. The aim of this research was to construct a machine learning model, using routinely collected farm data, to make probabilistic predictions at drying off for an individual cow's risk of a raised somatic cell count (hence intra-mammary infection) post-calving. Anonymized data were obtained as a large convenience sample from 108 UK dairy herds that undertook regular milk recording. The outcome measure evaluated was the presence of a raised somatic cell count in the 30 days post-calving in this observational study. Using a 56-farm training dataset, machine learning analysis was performed using the extreme gradient boosting decision tree algorithm, XGBoost. External validation was undertaken on a separate 28-farm test dataset. Statistical assessment to evaluate model performance using the external dataset returned calibration plots, a Scaled Brier Score of 0.095, and a Mean Absolute Calibration Error of 0.009. Test dataset model calibration performance indicated that the probability of a raised somatic cell count post-calving was well differentiated across probabilities to allow an end user to apply group-level risk decisions. Herd-level new intra-mammary infection rate during the dry period was a key driver of the probability that a cow had a raised SCC post-calving, highlighting the importance of optimizing environmental hygiene conditions. In conclusion, this research has determined that probabilistic classification of the risk of a raised SCC in the 30 days post-calving is achievable with a high degree of certainty, using routinely collected data. These predicted probabilities provide the opportunity for farmers to undertake risk decision-making by grouping cows based on their probabilities and optimizing management strategies for individual cows immediately after calving, according to their likelihood of intra-mammary infection.

Health Status Classification for Cows Using Machine Learning and Data Management on AWS Cloud

The health and welfare of livestock are significant for ensuring the sustainability and profitability of the agricultural industry. Addressing efficient ways to monitor and report the health status of individual cows is critical to prevent outbreaks and maintain herd productivity. The purpose of the study is to develop a machine learning (ML) model to classify the health status of milk cows into three categories. In this research, data are collected from existing non-invasive IoT devices and tools in a dairy farm, monitoring the micro- and macroenvironment of the cow in combination with particular information on age, days in milk, lactation, and more. A workflow of various data-processing methods is systematized and presented to create a complete, efficient, and reusable roadmap for data processing, modeling, and real-world integration. Following the proposed workflow, the data were treated, and five different ML algorithms were trained and tested to select the most descriptive one to monitor the health status of individual cows. The highest result for health status assessment is obtained by random forest classifier (RFC) with an accuracy of 0.959, recall of 0.954, and precision of 0.97. To increase the security, speed, and reliability of the work process, a cloud architecture of services is presented to integrate the trained model as an additional functionality in the Amazon Web Services (AWS) environment. The classification results of the ML model are visualized in a newly created interface in the client application.

CowMesh: a data-mesh architecture to unify dairy industry data for prediction and monitoring

Dairy is an economically significant industry that caters to the huge demand for food products in people's lives. To remain profitable, farmers need to manage their farms and the health of the dairy cows in their herds. There are, however, many risks to cow health that can lead to significant challenges to dairy farm management and have the potential to lead to significant losses. Such risks include cow udder infections (i.e., mastitis) and cow lameness. As automation and data recording become more common in the agricultural sector, dairy farms are generating increasing amounts of data. Recently, these data are being used to generate insights into farm and cow health, where the objective is to help farmers manage the health and welfare of dairy cows and reduce losses from cow health issues. Despite the level of data generation on dairy farms, this information is often difficult to access due to a lack of a single, central organization to collect data from individual farms. The prospect of such an organization, however, raises questions about data ownership, with some farmers reluctant to share their farm data for privacy reasons. In this study, we describe a new data mesh architecture designed for the dairy industry that focuses on facilitating access to data from farms in a decentralized fashion. This has the benefit of keeping the ownership of data with dairy farmers while bringing data together by providing a common and uniform set of protocols. Furthermore, this architecture will allow secure access to the data by research groups and product development groups, who can plug in new projects and applications built across the data. No similar framework currently exists in the dairy industry, and such a data mesh can help industry stakeholders by bringing the dairy farms of a country together in a decentralized fashion. This not only helps farmers, dairy researchers, and product builders but also facilitates an overview of all dairy farms which can help governments to decide on regulations to improve the dairy industry at a national level.

A Literature Review of Modeling Approaches Applied to Data Collected in Automatic Milking Systems

Automatic milking systems (AMS) have played a pioneering role in the advancement of Precision Livestock Farming, revolutionizing the dairy farming industry on a global scale. This review specifically targets papers that focus on the use of modeling approaches within the context of AMS. We conducted a thorough review of 60 articles that specifically address the topics of cows' health, production, and behavior/management Machine Learning (ML) emerged as the most widely used method, being present in 63% of the studies, followed by statistical analysis (14%), fuzzy algorithms (9%), deterministic models (7%), and detection algorithms (7%). A significant majority of the reviewed studies (82%) primarily focused on the detection of cows' health, with a specific emphasis on mastitis, while only 11% evaluated milk production. Accurate forecasting of dairy cow milk yield and understanding the deviation between expected and observed milk yields of individual cows can offer significant benefits in dairy cow management. Likewise, the study of cows' behavior and herd management in AMSs is under-explored (7%). Despite the growing utilization of machine learning (ML) techniques in the field of dairy cow management, there remains a lack of a robust methodology for their application. Specifically, we found a substantial disparity in adequately balancing the positive and negative classes within health prediction models.

Early detection of subclinical mastitis in lactating dairy cows using cow-level features

Subclinical mastitis in cows affects their health, well-being, longevity, and performance, leading to reduced productivity and profit. Early prediction of subclinical mastitis can enable dairy farmers to perform interventions to mitigate its effect. The present study investigated how well predictive models built using machine learning techniques can detect subclinical mastitis up to 7 d before its occurrence. The data set used consisted of 1,346,207 milk-day (i.e., a day when milk was collected on both morning and evening) records spanning 9 yr from 2,389 cows producing on 7 Irish research farms. Individual cow composite milk yield and maximum milk flow were available twice daily, whereas milk composition (i.e., fat, lactose, protein) and somatic cell count (SCC) were collected once per week. Other features describing parity, calving dates, predicted transmitting ability for SCC, body weight, and history of subclinical mastitis were also available. The results of the study showed that a gradient boosting machine model trained to predict the onset of subclinical mastitis 7 d before a subclinical case occurs achieved a sensitivity and specificity of 69.45 and 95.64%, respectively. Reduced data collection frequency, where milk composition and SCC were recorded only every 15, 30, 45, and 60 d was simulated by masking data, to reflect the frequency of recording of this data on commercial dairy farms in Ireland. The sensitivity and specificity scores reduced as recording frequency reduced with respective scores of 66.93 and 80.43% when milk composition and SCC were recorded just every 60 d. Results demonstrate that models built on data that could be recorded routinely available on commercial dairy farms, can achieve useful predictive ability of subclinical mastitis even with reduced frequency of milk composition and SCC recording.

Genomic and Phenotypic Udder Evaluation for Dairy Cattle Selection: A Review

The traditional point of view regarding dairy cattle selection has been challenged by recent genomic studies indicating that livestock productivity prediction can be redefined based on the evaluation of genomic and phenotypic data. Several studies that included different genomic-derived traits only indicated that interactions among them or even with conventional phenotypic evaluation criteria require further elucidation. Unfortunately, certain genomic and phenotypic-derived traits have been shown to be secondary factors influencing dairy production. Thus, these factors, as well as evaluation criteria, need to be defined. Owing to the variety of genomic and phenotypic udder-derived traits which may affect the modern dairy cow functionality and conformation, a definition of currently important traits in the broad sense is indicated. This is essential for cattle productivity and dairy sustainability. The main objective of the present review is to elucidate the possible relationships among genomic and phenotypic udder evaluation characteristics to define the most relevant traits related to selection for function and conformation in dairy cattle. This review aims to examine the potential impact of various udder-related evaluation criteria on dairy cattle productivity and explore how to mitigate the adverse effects of compromised udder conformation and functionality. Specifically, we will consider the implications for udder health, welfare, longevity, and production-derived traits. Subsequently, we will address several concerns covering the application of genomic and phenotypic evaluation criteria with emphasis on udder-related traits in dairy cattle selection as well as its evolution from origins to the present and future prospects.

Exploiting machine learning methods with monthly routine milk recording data and climatic information to predict subclinical mastitis in Italian Mediterranean buffaloes

Mastitis has detrimental effects on the world's dairy industry, reducing animal health, milk production and quality, as well as income for farmers. In addition, consumers' growing interest in food safety and rational usage of antibiotics highlights the need to develop novel strategies to improve mastitis detection, prevention, and management. In the present study we applied machine learning (ML) analyses to predict presence or absence of subclinical mastitis in Italian Mediterranean buffaloes, exploiting information collected the previous month during routine milk recording procedures, as well as climatic data. The data set included 3,891 records of 1,038 buffaloes from 6 herds located in Basilicata Region (South Italy). Prediction models were developed using 4 different ML algorithms (Generalized Linear Model, Support Vector Machines, Random Forest, and Neural Network) and 2 data set splitting approaches for the creation of the training and test sets (by record or by animal ID number, always with 80% of the data used for model training and the remaining 20% for model testing). Support Vector Machine was the best method to predict high or low somatic cell count at the subsequent test-day record in the validation set, and therefore it was used to estimate the contribution of each feature to the best model. Independently from the data set splitting approach, the most important features were somatic cell score, differential somatic cell count, electrical conductivity, and milk production. Among climatic data, the most informative were temperature and relative humidity. When the data were split by animal ID, an improvement in models' predictive performance on the test set was observed, suggesting this as the most appropriate data splitting approach in data sets with repeated measures to avoid data leakage. According to different metrics, Neural Network was the best method for making predictions on the test set. Our findings confirmed the promising role of ML methods to improve prevention and surveillance of subclinical mastitis, exploiting the large amount of data currently available to identify animals that would possibly have high somatic cell count the subsequent month.

A Study on the Uniform Distribution and Counting Method of Raw Cow's Milk Somatic Cells

The somatic cell count (SCC) in raw milk is an important basis for determining whether a cow is suffering from mastitis. To address the problem of an uneven distribution of somatic cells due to cell-adherent sedimentation, among other reasons, during milk sampling, which in turn results in unrepresentative somatic cell counting, a method is proposed for obtaining a uniform distribution of somatic cells and improving the counting accuracy based on a nine-cell grid microfluidic chip. Firstly, a simulation was performed to verify the uniformity of the somatic cell distribution within the chip observation cavities. Secondly, a nine-cell grid microfluidic chip was prepared and a negative-pressure injection system integrating staining and stirring was developed to ensure that the somatic cells were uniformly distributed and free from air contamination during the injection process. As well as the structure of the chip, a microscopic imaging system was developed, and the nine chip observation cavities were photographed. Finally, the somatic cells were counted and the uniformity of the somatic cell distribution was verified using image processing. The experimental results show that the standard deviation coefficient of the SCC in each group of nine images was less than 1.61%. The automatic counting accuracy of the system was between 97.07% and 99.47%. This research method lays the foundation for the detection and prevention of mastitis in cows.

Relationship between total and differential quarter somatic cell counts at dry-off and early lactation

Mastitis is a most common disease of dairy cows and causes tremendous economic loss to the dairy industry worldwide. Somatic cell counts (SCC) reflect the inflammatory response to infections and is a metric used as key indicator in mastitis screening programs, typically within the framework of national milk recording schemes. Besides the determination of total SCC, the differentiation of cell types has been described to be beneficial for a more definite description of the actual udder health status of dairy cows. Differential somatic cell count (DSCC) represents the combined proportion of polymorphonuclear leukocytes (PMN) and lymphocytes expressed as a percentage of the total. The aim of this study was to investigate the relationship between SCC and differential somatic cell count (DSCC) in individual quarter milk samples collected at different time points: at dry-off, after calving and at the lactation peak. We used individual quarter data from farms representing the specialized production system of Parmigiano Reggiano cheese in Northern Italy. Average DSCC values ranged between 44.9% and 56.3%, with higher values (60.4%-72.1%) in milk samples with ≥ 1 million SCC/ml (where the proportion of samples with DSCC > 70% can be as high as 0.73). Moderate overall correlations between DSCC and log(SCC) were estimated (Pearson = 0.42, Spearman = 0.38), with a clear increasing trend with parity and around the lactation peak (e.g. Pearson = 0.59 at 60 DIM in parity 4). Taking SCC values as indicators of subclinical mastitis, DSCC would diagnose mastitis with 0.75 accuracy. Data editing criteria do have an impact on results, with stricter filtering leading to lower correlations between log(SCC) and DSCC. In conclusion DSCC and SCC provide different descriptions of the udder health status of dairy cows which, at least to some extent, are independent. DSCC alone doesn't provide more accurate information than SCC at quarter level but, used in combination with SCC, can be of potential interest within the framework of milk recording programs, especially in the context of selective dry-cow therapy (SDCT). However, this needs further investigation and updated threshold values need to be selected and validated.

A Practical Application of Genomic Predictions for Mastitis Resistance in Italian Holstein Heifers

Heifers are a fundamental resource on farms, and their importance is reflected in both farm management and economy. Therefore, the selection of heifers to be reared on a farm should be carefully performed to select only the best animals. Genomic selection is available nowadays to evaluate animals in a fast and economic way. However, it is mainly used on the sire line and on performance traits. Ten farms were selected based on their 5-year records of average somatic cell count and evenly classified into high (>300,000 cells/mL) and low somatic cell count (<150,000 cells/mL). Genomic indexes (regarding both wellness and productive traits) were evaluated in 157 Italian Holstein heifers reared in the selected ten farms (90 from high-cells farms and 67 from low-cells ones). Linear mixed models were fitted to analyze the effects of the abovementioned genomic indexes on related phenotypes. Results have shown that farms classified into low somatic cell count had an overall better animal genomic pool compared to high somatic cell count ones. Additionally, the results shown in this study highlighted a difference in wellness genomic indexes in animals from farms with either a high or a low average somatic cell count. Applying genomic tools directly to heifer selection could improve economic aspects related to herd turnover.

Machine Learning Algorithms Highlight tRNA Information Content and Chargaff’s Second Parity Rule Score as Important Features in Discriminating Probiotics from Non-Probiotics

Simple Summary

Probiotics are a group of beneficial microorganisms that are symbionts of the human gut microbiome. The identification of new probiotics is therefore of paramount importance from both public health and commercial perspectives. In this study, we show for the first time that Artificial Intelligence algorithms can identify novel probiotics and also discriminate them from pathogenic organisms in the human gut. We were also able to determine the information content within tRNA sequences as the key genomic features capable of characterizing probiotics.

Abstract

Probiotic bacteria are microorganisms with beneficial effects on human health and are currently used in numerous food supplements. However, no selection process is able to effectively distinguish probiotics from non-probiotic organisms on the basis of their genomic characteristics. In the current study, four Machine Learning algorithms were employed to accurately identify probiotic bacteria based on their DNA characteristics. Although the prediction accuracies of all algorithms were excellent, the Neural Network returned the highest scores in all the evaluation metrics, managing to discriminate probiotics from non-probiotics with an accuracy greater than 90%. Interestingly, our analysis also highlighted the information content of the tRNA sequences as the most important feature in distinguishing the two groups of organisms probably because tRNAs have regulatory functions and might have allowed probiotics to evolve faster in the human gut environment. Through the methodology presented here, it was also possible to identify seven promising new probiotics that have a higher information content in their tRNA sequences compared to non-probiotics. In conclusion, we prove for the first time that Machine Learning methods can discriminate human probiotic from non-probiotic organisms underlining information within tRNA sequences as the most important genomic feature in distinguishing them.

Mastitis Control and Intramammary Antimicrobial Stewardship in Ireland: Challenges and Opportunities

The Veterinary Medicines Regulation (EU 2019/6) came into force in all EU member states on 28 January 2022. This regulation places particular emphasis on prudent and responsible antimicrobial use in food animal production. Key changes include restrictions on the prophylactic use of antimicrobials in animals, and the possibility to reserve certain antimicrobials for humans only. The Regulation presents challenges to the Irish dairy industry, particularly with respect to current approaches to dry cow therapy. In response, the CellCheck technical working group (TWG, a technical group working in support of CellCheck, the national mastitis control programme) have developed pragmatic national and farm-level recommendations in support of improved mastitis control and intramammary antimicrobial stewardship in the Irish dairy industry. This paper outlines these recommendations, and provides an overview of the evidence considered to inform the TWG during its work (including the Regulation, policy perspectives, international best-practice, international scientific reviews and specific Irish challenges). In many key areas of concern, the TWG recognises the challenges in seeking to shape recommendations in the absence of robust and practical scientific evidence. For this reason, some of the recommended actions are pragmatic in nature, informed by national and international experiences. Periodic programme review will be needed, informed by ongoing monitoring of key performance indicators, to identify those actions that are most effective in an Irish context.