Search

Dramatic improvements in measuring genetic variation across agriculturally relevant populations (genomics) must be matched by improvements in identifying and measuring relevant trait variation in such populations across many environments (phenomics). Identifying the most critical opportunities and challenges in genome to phenome (G2P) research is the focus of this paper. Previously (Genome Biol, 23(1):1-11, 2022), we laid out how Agricultural Genome to Phenome Initiative (AG2PI) will coordinate activities with USA federal government agencies expand public-private partnerships, and engage with external stakeholders to achieve a shared vision of future the AG2PI. Acting on this latter step, AG2PI organized the "Thinking Big: Visualizing the Future of AG2PI" two-day workshop held September 9-10, 2022, in Ames, Iowa, co-hosted with the United State Department of Agriculture's National Institute of Food and Agriculture (USDA NIFA). During the meeting, attendees were asked to use their experience and curiosity to review the current status of agricultural genome to phenome (AG2P) work and envision the future of the AG2P field. The topic summaries composing this paper are distilled from two 1.5-h small group discussions. Challenges and solutions identified across multiple topics at the workshop were explored. We end our discussion with a vision for the future of agricultural progress, identifying two areas of innovation needed: (1) innovate in genetic improvement methods development and evaluation and (2) innovate in agricultural research processes to solve societal problems. To address these needs, we then provide six specific goals that we recommend be implemented immediately in support of advancing AG2P research., (© 2023. The Author(s).)

In the United States, dairy calves are typically housed individually due to the perception of reduced risk of spreading infectious diseases between calves and the ability to monitor health on an individual calf basis. However, automated milk feeders (AMF) can provide individual monitoring of group-housed calves while allowing them to express more natural feeding behaviors and interact with each other. Research has shown that feeding behaviors recorded by AMF can be a helpful screening tool for detecting disease in dairy calves. Altogether, there is an opportunity to use the data from AMF to create a more robust and efficient model to predict disease, reducing the need for visual observation. Therefore, the objective of this observational study was to predict disease in preweaning dairy calves using AMF feeding behavior data and machine learning (ML) algorithms. This study was conducted on a dairy farm located in the Upper Midwest United States and visited weekly from July 2018 to May 2019. During farm visits, AMF data and calves' treatment records were collected, and calves were visually health-scored for attitude, ear position, ocular discharge, nasal discharge, hide dirtiness, and cough score. The final datasets used for the analyses consisted of 740 and 741 calves, with 1,007 (healthy = 594 and sick = 413) and 1,044 (healthy = 560 and sick = 484) observations (health events) for Data 1 and Data 2, respectively. Data 1 included only the weekly calf health scores observed by research personnel, whereas Data 2 included primarily daily calf treatment records by farm staff in addition to weekly health scores. Calf visit-level feeding behaviors from AMF data included milk intake (mL/d), drinking speed (mL/min), visit duration (min), rewarded (with milk being offered) and unrewarded (without milk) visits (number per d), and the interval between visits (min). Three approaches were used to predict health status: generalized linear model, random forest, and gradient boosting machine. A total of 16 models were built using different combinations of behavior parameters, including the number of rewarded visits, number of unrewarded visits, visit duration, the interval between visits, intake, intake divided by rewarded visits, drinking speed, and drinking speed divided by rewarded visits, and also calf age at the sick day as predictor variables. Of all algorithms, random forest and gradient boosting had the best performance predicting the health status of dairy calves. The results indicated that weekly health scores were not enough to predict calf health status. However, daily treatment records and AMF data were sufficient for creating predictive algorithms (e.g., F1-scores of 0.775 and 0.784 for random forest and gradient boosting, Data 2). This study suggests that ML was effective in determining the specific visit-level feeding behaviors that can be used to predict disease in group-housed preweaning dairy calves. Implementing these ML algorithms could reduce the need for visual calf observation on farms, minimizing labor time and improving calf health., (© 2024, The Authors. Published by Elsevier Inc. and Fass Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).)

Respiratory rate (RR) is commonly employed for identifying animals experiencing heat-stress conditions and respiratory diseases. Recent advancements in computer vision algorithms have enabled the estimation of the RR in dairy cows through image-based approaches, with a primary focus on standing positions, thermal imaging, and deep learning techniques. In this study, our objective was to develop a system capable of accurately predicting the RR of lying Holstein cows under unrestrained conditions using red, green, and blue (RGB) and infrared (IR) night vision images. Thirty lactating cows were continuously recorded for 12 h per day over a 3-d period, capturing at least one 30-s video segment of each cow during lying time. A total of 95 videos were manually annotated with rectangular bounding boxes encompassing the flank area (region of interest; ROI) of the lying cows. For future applications, we trained a model for ROI identification using YOLOv8 to avoid manual annotations. The observed RR was determined by visual counting of breaths in each video. To predict the RR, we devised an image processing pipeline involving (1) capturing the ROI for the entire video, (2) reshaping the pixel intensity of each image channel into a 2-dimensional object and calculating its per-frame mean, (3) applying fast Fourier transform (FFT) to the average pixel intensity vector, (4) filtering frequencies specifically associated with respiratory movements, and (5) executing inverse FFT on the denoized data and identifying peaks on the resulting plot, with the count of peaks serving as the predicted RR per minute. The evaluation metrics, root mean squared error of prediction (RMSEP) and R 2 , yielded values of 8.3 breaths/min (17.1% of the mean RR) and 0.77, respectively. To further validate the method, an additional dataset comprising preweaning dairy calves was used, consisting of 42 observations from 25 calves. The RMSEP and R 2 values for this dataset were 13.0 breaths/min and 0.73, respectively. The model trained to identify the ROI exhibited a precision of 100%, a recall of 71.8%, and an F 1 score of 83.6% for bounding box detection. These are promising results for the implementation of this pipeline in future studies. The application of FFT to signals acquired from both RGB and IR images proved to be an effective and accurate method for computing the RR of cows in unrestrained conditions., (© 2023.)

Haemonchus contortus is the most pathogenic blood-feeding parasitic in sheep, causing anemia and consequently changes in the color of the ocular conjunctiva, from the deep red of healthy sheep to shades of pink to practically white of non-healthy sheep. In this context, the Famacha method has been created for detecting sheep unable to cope with the infection by H. contortus , through visual assessment of ocular conjunctiva coloration. Thus, the objectives of this study were (1) to extract ocular conjunctiva image features to automatically classify Famacha score and compare two classification models (multinomial logistic regression-MLR and random forest-RF) and (2) to evaluate the applicability of the best classification model on three sheep farms. The dataset consisted of 1,156 ocular conjunctiva images from 422 animals. RF model was used to segment the images, i.e., to select the pixels that belong to the ocular conjunctiva. After segmentation, the quantiles (1%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 99%) of color intensity in each image channel (red, blue, and green) were determined and used as explanatory variables in the classification models, and the Famacha scores 1 (non-anemic) to 5 (severely anemic) were the target classes to be predicted (scores 1 to 5, with 162, 255, 443, 266, and 30 images, respectively). For objective 1, the performance metrics (precision and sensitivity) were obtained using MLR and RF models considering data from all farms randomly split. For objective 2, a leave-one-farm-out cross-validation technique was used to assess prediction quality across three farms (farms A, B, and C, with 726, 205, and 225 images, respectively). The RF provided the best performances in predicting anemic animals, as indicated by the high values of sensitivity for Famacha score 3 (80.9%), 4 (46.2%), and 5 (60%) compared to the MLR model. The precision of the RF was 72.7% for Famacha score 1 and 62.5% for Famacha score 2. These results indicate that is possible to successfully predict Famacha score, especially for scores 2 to 4, in sheep via image analysis and RF model using ocular conjunctiva images collected in farm conditions. As expected, model validation excluding entire farms in cross-validation presented a lower prediction quality. Nonetheless, this setup is closer to reality because the developed models are supposed to be used across farms, including new ones, and with different environments and management conditions., Competing Interests: There are no potential conflicts of interest that may affect our ability to objectively present or review the research data submitted here., (Published by Oxford University Press for the American Society of Animal Science 2023.)

Contemporary approaches for animal identification use deep learning techniques to recognize coat color patterns and identify individual animals in a herd. However, deep learning algorithms usually require a large number of labeled images to achieve satisfactory performance, which creates the need to manually label all images when automated methods are not available. In this study, we evaluated the potential of a semi-supervised learning technique called pseudo-labeling to improve the predictive performance of deep neural networks trained to identify Holstein cows using labeled training sets of varied sizes and a larger unlabeled dataset. By using such technique to automatically label previously unlabeled images, we observed an increase in accuracy of up to 20.4 percentage points compared to using only manually labeled images for training. Our final best model achieved an accuracy of 92.7% on an independent testing set to correctly identify individuals in a herd of 59 cows. These results indicate that it is possible to achieve better performing deep neural networks by using images that are automatically labeled based on a small dataset of manually labeled images using a relatively simple technique. Such strategy can save time and resources that would otherwise be used for labeling, and leverage well annotated small datasets., (© 2023. Springer Nature Limited.)

Automated milk feeders (AMF) are an attractive option for producers interested in adopting practices that offer greater behavioral freedom for calves and can potentially improve labor management. These feeders give farmers the opportunity to have a more flexible labor schedule and more efficiently feed group-housed calves. However, housing calves in group systems can pose challenges for monitoring calf health on an individual basis, potentially leading to increased morbidity and mortality. Feeding behavior recorded by AMF software could potentially be used as an indicator of disease. Therefore, the objective of this observational study was to investigate the association between feeding behaviors and disease in preweaning group-housed dairy calves fed with AMF. The study was conducted at a dairy farm located in the Upper Midwest United States and included a final data set of 599 Holstein heifer calves. The farm was visited on a weekly basis from May 2018, to May 2019, when calves were visually health scored and AMF data were collected. Calf health scores included calf attitude, ear position, ocular discharge, nasal discharge, hide dirtiness, cough score, and rectal temperatures. Generalized additive mixed models (GAMM) were used to identify associations between feeding behavior and disease. The final quasibinomial GAMM included the fixed (main and interactions) effects of feeding behavior at calf visit-level including milk intake (mL/d), drinking speed (mL/min), visit duration (min), rewarded (with milk being offered) and unrewarded (without milk) visits (number per day), and interval between visits (min), as well as the random effects of calf age in regard to their relationship with calf health status. Total milk intake (mL/d), drinking speed (mL/min), interval between visits (min) to the AMF, calf age (d), and rewarded visits were significantly associated with dairy calf health status. These results indicate that as total milk intake and drinking speed increased, the risk of calves being sick decreased. In contrast, as the interval between visits and age increased, the risk of calves being sick also increased. This study suggests that AMF data may be a useful screening tool for detecting disease in dairy calves. In addition, GAMM were shown to be a simple and flexible approach to modeling calf health status, as they can cope with non-normal data distribution of the response variable, capture nonlinear relationships between explanatory and response variables and accommodate random effects., (The Authors. Published by Elsevier Inc. and Fass Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).)

Computer vision systems have emerged as a potential tool to monitor the behavior of livestock animals. Such high-throughput systems can generate massive redundant data sets for training and inference, which can lead to higher computational and economic costs. The objectives of this study were (1) to develop a computer vision system to individually monitor detailed feeding behaviors of group-housed dairy heifers, and (2) to determine the optimal frequency of image acquisition to perform inference with minimal effect on feeding behavior prediction quality. Eight Holstein heifers (96 ± 6 d old) were housed in a group and a total of 25,214 images (1 image every second) were acquired using 1 RGB camera. A total of 2,209 images were selected and each animal in the image was labeled with its respective identification (1-8). The label was annotated only on animals that were at the feed bunk (head through the feed rail). From the labeled images, 1,392 were randomly selected to train a deep learning algorithm for object detection with YOLOv3 ("You Only Look Once" version 3) and 154 images were used for validation. An independent data set (testing set = 663 out of the 2,209 images) was used to test the algorithm. The average accuracy for identifying individual animals in the testing set was 96.0%, and for each individual heifer from 1 to 8 the accuracy was 99.2, 99.6, 99.2, 99.6, 99.6, 99.2, 99.4, and 99.6%, respectively. After identifying the animals at the feed bunk, we computed the following feeding behavior parameters: number of visits (NV), mean visit duration (MVD), mean interval between visits (MIBV), and feeding time (FT) for each heifer using a data set composed by 8,883 sequential images (1 image every second) from 4 time points. The coefficient of determination (R 2 ) was 0.39, 0.78, 0.48, and 0.99, and the root mean square error (RMSE) were 12.3 (count), 0.78, 0.63, and 0.31 min for NV, MVD, MIBV, and FT, respectively, considering 1 image every second. When we moved from 1 image per second to 1 image every 5 (MIBV) or 10 (NV, MDV, and FT) s, the R 2 observed were 0.55 (NV), 0.74 (MVD), 0.70 (MIBV), and 0.99 (FT); and the RMSE were 2.27 (NV, count), 0.38 min (MVD), 0.22 min (MIBV), and 0.44 min (FT). Our results indicate that computer vision systems can be used to individually identify group-housed Holstein heifers (overall accuracy = 99.4%). Based on individual identification, feeding behavior such as MVD, MIBV, and FT can be monitored with reasonable accuracy and precision. Regardless of the frequency for optimal image acquisition, our results suggested that longer time intervals of image acquisition would reduce data collecting and model inference while maintaining adequate predictive performance. However, we did not find an optimal time interval for all feeding behavior; instead, the optimal frequency of image acquisition is phenotype-specific. Overall, the best R 2 and RMSE for NV, MDV, and FT were achieved using 1 image every 10 s, and for MIBV it was achieved using 1 image every 5 s, and in both cases model inference and data storage could be drastically reduced., (The Authors. Published by Elsevier Inc. and Fass Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).)

Our aim was to explore whether changes in plasma essential AA (EAA) concentration ([EAA] p ) or profile (defined here as the molar proportion of individual [EAA] p relative to the total [EAA] p ) may serve as an indicator of the EAA status of a cow. We undertook a meta-analysis with the objectives to determine if different plasma EAA profiles exist among cows and to explore the association of [EAA] p or the profile of EAA with lactating cow performance and measures of N utilization. We hypothesized the existence of differences in [EAA] p and different plasma EAA profile for cows with greater milk output, feed efficiency, and greater N use efficiency (NUE; milk true protein-N:N intake) compared with cows with lower milk output, feed efficiency, and lower NUE. The data set included 22 feeding trials and 96 dietary treatments. First, a mixed-effect model analysis was used to predict [EAA] p in response to the categorical fixed effect of EAA, continuous fixed effect of National Research Council model-predicted metabolizable protein (MP) supply, continuous fixed effect of body weight, the fixed effect of EAA and MP supply interaction, the fixed effect of EAA and body weight interaction, and the random effect of study. Then, residuals of the model were standardized based on Z-score and clustered using the hierarchical method (Euclidean distance and Ward's minimum variance method) resulting in 2 clusters. Finally, a fixed-effect model was used to evaluate the significance with which clusters were associated with [EAA] p , cow performance, feed efficiency, and NUE. The total concentration of [EAA] p was lower (784 vs. 983 µM) and the concentration of each EAA was on average 22 µM lower for cows in cluster 1 compared with cluster 2 with the smallest and greatest difference found for Met (4 µM) and Val (59 µM), respectively. The percentage difference in [EAA] p was the smallest for Thr (-5.3%) and the greatest for Leu (-37.1%). There was no difference between clusters for Arg, His, and Met molar proportions; however, cows in cluster 1 had a lower molar proportion of Leu and a tendency for lower molar proportion of Val compared with cows in cluster 2. Additionally, cows in cluster 1 had greater molar proportions of Ile, Lys, and Thr and a tendency for greater molar proportion of Phe compared with cows in cluster 2. The fixed-effect model analysis indicated that cows in cluster 1 had higher milk energy output (+3.2 Mcal/d), true protein yield (+87 g/d) and fat yield (+236 g/d), feed efficiency (milk Mcal:dry matter intake; +8% unit), and a tendency for greater MP efficiency (Milk true protein/MP supply; +2.3% unit) than cows in cluster 2. These results suggested greater use of EAA by the mammary gland (as reflected by greater milk protein synthesis) and lower hepatic catabolism of AA (as reflected by a tendency to greater MP efficiency) in cows of cluster 1 compared with cluster 2. Our findings should be evaluated further, including whether the relative molar proportions of plasma EAA might serve as a holistic indicator of the EAA status of cows as related to their productivity, feed efficiency and N utilization., (© 2022, The Authors. Published by Elsevier Inc. and Fass Inc. on behalf of the American Dairy Science Association®. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).)

The use of precision farming technologies, such as milking robots, automated calf feeders, wearable sensors, and others, has significantly increased in dairy operations over the last few years. The growing interest in farming technologies to reduce labor, maximize productivity, and increase profitability is becoming noticeable in several countries, including Brazil. Information regarding technology adoption, perception, and effectiveness in dairy farms could shed light on challenges that need to be addressed by scientific research and extension programs. The objective of this study was to characterize Brazilian dairy farms based on technology usage. Factors such as willingness to invest in precision technologies, adoption of sensor systems, farmer profile, farm characteristics, and production indexes were investigated in 378 dairy farms located in Brazil. A survey with 22 questions was developed and distributed via Google Forms from July 2018 to July 2020. The farms were then classified into seven clusters: (1) top yield farms; (2) medium-high yield, medium-tech; (3) medium yield and top high-tech; (4) medium yield and medium-tech; (5) young medium-low yield and low-tech; (6) elderly medium-low yield and low-tech; and (7) low-tech grazing. The most frequent technologies adopted by producers were milk meters systems (31.7%), milking parlor smart gate (14.5%), sensor systems to detect mastitis (8.4%), cow activity meter (7.1%), and body temperature (7.9%). Based on a scale containing numerical values (1-5), producers indicated "available technical support" (mean; σ 2 ) (4.55; 0.80) as the most important decision criterion involved in adopting technology, followed by "return on investment-ROI" (4.48; 0.80), "user-friendliness" (4.39; 0.88), "upfront investment cost" (4.36; 0.81), and "compatibility with farm management software" (4.2; 1.02). The most important factors precluding investment in precision dairy technologies were the need for investment in other sectors of the farm (36%), the uncertainty of ROI (24%), and lack of integration with other farm systems and software (11%). Farmers indicated that the most useful technologies were automatic milk meters systems (mean; σ 2 ) (4.05; 1.66), sensor systems for mastitis detection (4.00; 1.57), automatic feeding systems (3.50; 2.05), cow activity meter (3.45; 1.95), and in-line milk analyzers (3.45; 1.95). Overall, the concerns related to data integration, ROI, and user-friendliness of technologies are similar to those of dairy farms located in other countries. Increasing available technical support for sensing technology can have a positive impact on technology adoption.

Wearable sensors have been explored as an alternative for real-time monitoring of cattle feeding behavior in grazing systems. To evaluate the performance of predictive models such as machine learning (ML) techniques, data cross-validation (CV) approaches are often employed. However, due to data dependencies and confounding effects, poorly performed validation strategies may significantly inflate the prediction quality. In this context, our objective was to evaluate the effect of different CV strategies on the prediction of grazing activities in cattle using wearable sensor (accelerometer) data and ML algorithms. Six Nellore bulls (average live weight of 345 ± 21 kg) had their behavior visually classified as grazing or not-grazing for a period of 15 d. Elastic Net Generalized Linear Model (GLM), Random Forest (RF), and Artificial Neural Network (ANN) were employed to predict grazing activity (grazing or not-grazing) using 3-axis accelerometer data. For each analytical method, three CV strategies were evaluated: holdout, leave-one-animal-out (LOAO), and leave-one-day-out (LODO). Algorithms were trained using similar dataset sizes (holdout: n = 57,862; LOAO: n = 56,786; LODO: n = 56,672). Overall, GLM delivered the worst prediction accuracy (53%) compared with the ML techniques (65% for both RF and ANN), and ANN performed slightly better than RF for LOAO (73%) and LODO (64%) across CV strategies. The holdout yielded the highest nominal accuracy values for all three ML approaches (GLM: 59%, RF: 76%, and ANN: 74%), followed by LODO (GLM: 49%, RF: 61%, and ANN: 63%) and LOAO (GLM: 52%, RF: 57%, and ANN: 57%). With a larger dataset (i.e., more animals and grazing management scenarios), it is expected that accuracy could be increased. Most importantly, the greater prediction accuracy observed for holdout CV may simply indicate a lack of data independence and the presence of carry-over effects from animals and grazing management. Our results suggest that generalizing predictive models to unknown (not used for training) animals or grazing management may incur poor prediction quality. The results highlight the need for using management knowledge to define the validation strategy that is closer to the real-life situation, i.e., the intended application of the predictive model., (© The Author(s) 2021. Published by Oxford University Press on behalf of the American Society of Animal Science. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Predicting dry matter intake (DMI) and feed efficiency by leveraging the use of data streams available on farm could aid efforts to improve the feed efficiency of dairy cattle. Residual feed intake (RFI) is the difference between predicted and observed feed intake after accounting for body size, body weight change, and milk production, making it a valuable metric for feed efficiency research. Our objective was to develop and evaluate DMI and RFI prediction models using multiple linear regression (MLR), partial least squares regression, artificial neural networks, and stacked ensembles using different combinations of cow descriptive, performance, sensor-derived behavioral (SMARTBOW; Zoetis), and blood metabolite data. Data were collected from mid-lactation Holstein cows (n = 124; 102 multiparous, 22 primiparous) split equally between 2 replicates of 45-d duration with ad libitum access to feed. Within each predictive approach, 4 data streams were added in sequence: dataset M (week of lactation, parity, milk yield, and milk components), dataset MB (dataset M plus body condition score and metabolic body weight), dataset MBS (dataset MB plus sensor-derived behavioral variables), and dataset MBSP (dataset MBS plus physiological blood metabolites). The combination of 4 datasets and 4 analytical approaches resulted in 16 analyses of DMI and RFI, using variables averaged within cow across the study period. Additional models using weekly averaged data within cow and study were built using all predictive approaches for datasets M, MB, and MBS. Model performance was assessed using the coefficient of determination, concordance correlation coefficient, and root mean square error of prediction. Predictive models of DMI performed similarly across all approaches, and models using dataset MBS had the greatest model performance. The best approach-dataset combination was MLR-dataset MBS, although several models performed similarly. Weekly DMI models had the greatest performance with MLR and partial least squares regression approaches. Dataset MBS models had incrementally better performance than datasets MB and M. Within each approach-dataset combination, models with DMI averaged over the study period had slightly greater model performance than DMI averaged weekly. Predictive performance of all RFI models was poor, but slight improvements when using MLR applied to dataset MBS suggest that rumination and activity behaviors may explain some of the variation in RFI. Overall, similar performance of MLR, compared with machine learning techniques, indicates MLR may be sufficient to predict DMI. The improvement in model performance with each additional data stream supports the idea of integrating data streams to improve model predictions and farm management decisions., (Copyright © 2021 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)

Background: Deep neural networks (DNN) are a particular case of artificial neural networks (ANN) composed by multiple hidden layers, and have recently gained attention in genome-enabled prediction of complex traits. Yet, few studies in genome-enabled prediction have assessed the performance of DNN compared to traditional regression models. Strikingly, no clear superiority of DNN has been reported so far, and results seem highly dependent on the species and traits of application. Nevertheless, the relatively small datasets used in previous studies, most with fewer than 5000 observations may have precluded the full potential of DNN. Therefore, the objective of this study was to investigate the impact of the dataset sample size on the performance of DNN compared to Bayesian regression models for genome-enable prediction of body weight in broilers by sub-sampling 63,526 observations of the training set., Results: Predictive performance of DNN improved as sample size increased, reaching a plateau at about 0.32 of prediction correlation when 60% of the entire training set size was used (i.e., 39,510 observations). Interestingly, DNN showed superior prediction correlation using up to 3% of training set, but poorer prediction correlation after that compared to Bayesian Ridge Regression (BRR) and Bayes Cπ. Regardless of the amount of data used to train the predictive machines, DNN displayed the lowest mean square error of prediction compared to all other approaches. The predictive bias was lower for DNN compared to Bayesian models, across all dataset sizes, with estimates close to one with larger sample sizes., Conclusions: DNN had worse prediction correlation compared to BRR and Bayes Cπ, but improved mean square error of prediction and bias relative to both Bayesian models for genome-enabled prediction of body weight in broilers. Such findings, highlights advantages and disadvantages between predictive approaches depending on the criterion used for comparison. Furthermore, the inclusion of more data per se is not a guarantee for the DNN to outperform the Bayesian regression methods commonly used for genome-enabled prediction. Nonetheless, further analysis is necessary to detect scenarios where DNN can clearly outperform Bayesian benchmark models.

Computer Vision, Digital Image Processing, and Digital Image Analysis can be viewed as an amalgam of terms that very often are used to describe similar processes. Most of this confusion arises because these are interconnected fields that emerged with the development of digital image acquisition. Thus, there is a need to understand the connection between these fields, how a digital image is formed, and the differences regarding the many sensors available, each best suited for different applications. From the advent of the charge-coupled devices demarking the birth of digital imaging, the field has advanced quite fast. Sensors have evolved from grayscale to color with increasingly higher resolution and better performance. Also, many other sensors have appeared, such as infrared cameras, stereo imaging, time of flight sensors, satellite, and hyperspectral imaging. There are also images generated by other signals, such as sound (ultrasound scanners and sonars) and radiation (standard x-ray and computed tomography), which are widely used to produce medical images. In animal and veterinary sciences, these sensors have been used in many applications, mostly under experimental conditions and with just some applications yet developed on commercial farms. Such applications can range from the assessment of beef cuts composition to live animal identification, tracking, behavior monitoring, and measurement of phenotypes of interest, such as body weight, condition score, and gait. Computer vision systems (CVS) have the potential to be used in precision livestock farming and high-throughput phenotyping applications. We believe that the constant measurement of traits through CVS can reduce management costs and optimize decision-making in livestock operations, in addition to opening new possibilities in selective breeding. Applications of CSV are currently a growing research area and there are already commercial products available. However, there are still challenges that demand research for the successful development of autonomous solutions capable of delivering critical information. This review intends to present significant developments that have been made in CVS applications in animal and veterinary sciences and to highlight areas in which further research is still needed before full deployment of CVS in breeding programs and commercial farms., (Copyright © 2020 Fernandes, Dórea and Rosa.)

High-throughput phenotyping technologies are growing in importance in livestock systems due to their ability to generate real-time, non-invasive, and accurate animal-level information. Collecting such individual-level information can generate novel traits and potentially improve animal selection and management decisions in livestock operations. One of the most relevant tools used in the dairy and beef industry to predict complex traits is infrared spectrometry, which is based on the analysis of the interaction between electromagnetic radiation and matter. The infrared electromagnetic radiation spans an enormous range of wavelengths and frequencies known as the electromagnetic spectrum. The spectrum is divided into different regions, with near- and mid-infrared regions being the main spectral regions used in livestock applications. The advantage of using infrared spectrometry includes speed, non-destructive measurement, and great potential for on-line analysis. This paper aims to review the use of mid- and near-infrared spectrometry techniques as tools to predict complex dairy and beef phenotypes, such as milk composition, feed efficiency, methane emission, fertility, energy balance, health status, and meat quality traits. Although several research studies have used these technologies to predict a wide range of phenotypes, most of them are based on Partial Least Squares (PLS) and did not considered other machine learning (ML) techniques to improve prediction quality. Therefore, we will discuss the role of analytical methods employed on spectral data to improve the predictive ability for complex traits in livestock operations. Furthermore, we will discuss different approaches to reduce data dimensionality and the impact of validation strategies on predictive quality., (Copyright © 2020 Bresolin and Dórea.)

Computer vision systems (CVS) have been shown to be a powerful tool for the measurement of live pig body weight (BW) with no animal stress. With advances in precision farming, it is now possible to evaluate the growth performance of individual pigs more accurately. However, important traits such as muscle and fat deposition can still be evaluated only via ultrasound, computed tomography, or dual-energy x-ray absorptiometry. Therefore, the objectives of this study were: 1) to develop a CVS for prediction of live BW, muscle depth (MD), and back fat (BF) from top view 3D images of finishing pigs and 2) to compare the predictive ability of different approaches, such as traditional multiple linear regression, partial least squares, and machine learning techniques, including elastic networks, artificial neural networks, and deep learning (DL). A dataset containing over 12,000 images from 557 finishing pigs (average BW of 120 ± 12 kg) was split into training and testing sets using a 5-fold cross-validation (CV) technique so that 80% and 20% of the dataset were used for training and testing in each fold. Several image features, such as volume, area, length, widths, heights, polar image descriptors, and polar Fourier transforms, were extracted from the images and used as predictor variables in the different approaches evaluated. In addition, DL image encoders that take raw 3D images as input were also tested. This latter method achieved the best overall performance, with the lowest mean absolute scaled error (MASE) and root mean square error for all traits, and the highest predictive squared correlation (R2). The median predicted MASE achieved by this method was 2.69, 5.02, and 13.56, and R2 of 0.86, 0.50, and 0.45, for BW, MD, and BF, respectively. In conclusion, it was demonstrated that it is possible to successfully predict BW, MD, and BF via CVS on a fully automated setting using 3D images collected in farm conditions. Moreover, DL algorithms simplified and optimized the data analytics workflow, with raw 3D images used as direct inputs, without requiring prior image processing., (© The Author(s) 2020. Published by Oxford University Press on behalf of the American Society of Animal Science. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

The objective of this study was to investigate the effects of energy supplementation and pre-grazing sward height on grazing behavior, nutrient intake, digestion, and metabolism of cattle in tropical pastures managed as a rotational grazing system. Eight rumen-cannulated Nellore steers (24 mo of age; 300 ± 6.0 kg body weight [BW]) were used in a replicated 4 × 4 Latin square design. Treatments consisted of two levels of energy supplementation (0% [none] or 0.3% of BW of ground corn on an as-fed basis) and two pre-grazing sward heights (25 cm [defined by 95% light interception (LI)] or 35 cm [defined by ≥ 97.5% LI]) constituting four treatments. Steers grazed Marandu Palisadegrass [Brachiaria brizantha Stapf. cv. Marandu] and post-grazing sward height was 15 cm for all treatments. Forage dry matter intake (DMI) was increased (P = 0.01) when sward height was 25 cm (1.86% vs. 1.32% BW) and decreased (P = 0.04) when 0.3% BW supplement was fed (1.79% vs. 1.38% BW). Total and digestible DMI were not affected by energy supplementation (P = 0.57) but were increased when sward height was 25 cm (P = 0.01). Steers grazing the 25-cm sward height treatment spent less time grazing and more time resting, took fewer steps between feeding stations, and had a greater bite rate compared with 35-cm height treatment (P < 0.05). Energy supplementation reduced grazing time (P = 0.02) but did not affect any other grazing behavior parameter (P = 0.11). Energy supplementation increased (P < 0.01) diet dry matter digestibility but had no effect on crude protein and neutral detergent fiber digestibilities (P = 0.13). Compared with 35-cm pre-grazing sward height, steers at 25 cm presented lower rumen pH (6.39 vs. 6.52) and greater rumen ammonia nitrogen (11.22 vs. 9.77 mg/dL) and N retention (49.7% vs. 20.8%, P < 0.05). The pre-grazing sward height of 25 cm improved harvesting efficiency and energy intake by cattle, while feeding 0.3% of BW energy supplement did not increase the energy intake of cattle on tropical pasture under rotational grazing., (© The Author(s) 2020. Published by Oxford University Press on behalf of the American Society of Animal Science. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

With agriculture rapidly becoming a data-driven field, it is imperative to extract useful information from large data collections to optimize the production systems. We compared the efficacy of regression (linear regression or generalized linear regression [GLR] for continuous or categorical outcomes, respectively), random forests (RF) and multilayer neural networks (NN) to predict beef carcass weight (CW), age when finished (AS), fat deposition (FD), and carcass quality (CQ). The data analyzed contained information on over 4 million beef cattle from 5,204 farms, corresponding to 4.3% of Brazil's national production between 2014 and 2016. Explanatory variables were integrated from different data sources and encompassed animal traits, participation in a technical advising program, nutritional products sold to farms, economic variables related to beef production, month when finished, soil fertility, and climate in the location in which animals were raised. The training set was composed of information collected in 2014 and 2015, while the testing set had information recorded in 2016. After parameter tuning for each algorithm, models were used to predict the testing set. The best model to predict CW and AS was RF (CW: predicted root mean square error = 0.65, R2 = 0.61, and mean absolute error = 0.49; AS: accuracy = 28.7%, Cohen's kappa coefficient [Kappa] = 0.08). While the best approach for FD and CQ was GLR (accuracy = 45.7%, Kappa = 0.05, and accuracy = 58.7%, Kappa = 0.09, respectively). Across all models, there was a tendency for better performance with RF and regression and worse with NN. Animal category, nutritional plan, cattle sales price, participation in a technical advising program, and climate and soil in which animals were raised were deemed important for prediction of meat production and quality with regression and RF. The development of strategies for prediction of livestock production using real-world large-scale data will be core to projecting future trends and optimizing the allocation of resources at all levels of the production chain, rendering animal production more sustainable. Despite beef cattle production being a complex system, this analysis shows that by integrating different sources of data it is possible to forecast meat production and quality at the national level with moderate-high levels of accuracy., (© The Author(s) 2020. Published by Oxford University Press on behalf of the American Society of Animal Science. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

Computer vision applications in livestock are appealing since they enable measurement of traits of interest without the need to directly interact with the animals. This allows the possibility of multiple measurements of traits of interest with minimal animal stress. In the current study, an automated computer vision system was devised and evaluated for extraction of features of interest, as body measurements and shape descriptors, and prediction of body weight in pigs. From the 655 pigs that had data collected 580 had more than 5 frames recorded and were used for development of the predictive models. The cross-validation for the models developed with data from nursery and finishing pigs presented an R2 ranging from 0.86 (random selected image) to 0.94 (median of images truncated on the third quartile), whereas with the dataset without nursery pigs, the R2 estimates ranged from 0.70 (random selected image) to 0.84 (median of images truncated on the third quartile). However, overall the mean absolute error was lower for the models fitted without data on nursery animals. From the body measures extracted from the image, body volume, area, and length were the most informative for prediction of body weight. The inclusion of the remaining body measurements (width and heights) or shape descriptors to the model promoted significant improvement of the predictions, whereas the further inclusion of sex and line effects were not significant.

Two experiments were conducted to evaluate the performance responses of finishing feedlot cattle to dietary addition of essential oils and exogenous enzymes. The treatments in each experiment consisted of (DM basis): MON-sodium monensin (26 mg/kg); BEO-a blend of essential oils (90 mg/kg); BEO+MON-a blend of essential oils plus monensin (90 mg/kg + 26 mg/kg, respectively); BEO+AM-a blend of essential oils plus exogenous α-amylase (90 mg/kg + 560 mg/kg, respectively); and BEO+AM+PRO-a blend of essential oils plus exogenous α-amylase and exogenous protease (90 mg/kg + 560 mg/kg + 840 mg/kg, respectively). Exp. 1 consisted of a 93-d finishing period using 300 Nellore bulls in a randomized complete block design. Animals fed BEO had higher DMI (P < 0.001) but similar feed efficiency to animals fed MON (P ≥ 0.98). Compared with MON, the combination of BEO+AM resulted in 810 g greater DMI (P = 0.001), 190 g greater average daily gain (P = 0.04), 18 kg heavier final body weight (P = 0.04), and 12 kg heavier hot carcass weight (P = 0.02), although feed efficiency was not significantly different between BEO+AM and MON (P = 0.89). Combining BEO+MON tended to decrease hot carcass weight compared with BEO alone (P = 0.08) but not compared with MON (P = 0.98). Treatments did not impact observed dietary net energy values (P ≥ 0.74) or the observed:expected net energy ratio (P ≥ 0.11). In Exp. 2, five ruminally cannulated Nellore steers were used to evaluate intake, apparent total tract digestibility of nutrients, and ruminal parameters in a 5 × 5 Latin square design. Feeding BEO increased the total tract digestibility of CP compared to MON (P = 0.03). Compared to MON, feeding the combination of BEO+MON increased the intake of CP (P = 0.04) and NDF (P = 0.05), with no effects on total tract digestibility of nutrients (P ≥ 0.56), except for a tendency (P = 0.09) to increase CP digestibility. Intakes of all nutrients measured, except for ether extract (P = 0.16) were greater in animals fed BEO+AM when compared with MON (P ≤ 0.03), with no differences on total tract nutrient digestibilities (P ≥ 0.11) between these two treatments. In summary, diets containing the BEO used herein enhanced DMI of growing-finishing feedlot cattle compared with a basal diet containing MON without impair feed efficiency. A synergism between BEO and AM was detected, further increasing cattle performance and carcass production compared to MON.

This study compared dry matter (DM) predictions of 3 handheld near-infrared spectrophotometer (NIRS) units (Moisture Tracker, Digi-Star Inc., Fort Atkinson, WI) to conventional oven drying at 60°C using 2 alfalfa and 2 corn silages. In addition, on-farm DM methods [microwave, Koster tester (Koster Moisture Tester Inc., Brunswick, OH), and food dehydrator methods] were also compared. Corn and alfalfa silages (1,600 g) obtained from the University of Wisconsin Dairy Cattle Center (DCC) and the Arlington Research Station (ARS) were analyzed for DM daily for 20 d. Two NIRS calibration methods were also tested within each unit. The DM predicted from the factory-preset calibrations was NIR f . The adjusted DM prediction was NIR a , where the average difference between oven-dried and NIR f determined on duplicate forage samples for 3 d before the experiment was used as a bias adjustment for all subsequent DM determinations. The average predicted DM from the 20 scans was recorded as the forage DM. The process was repeated 3 times with each NIRS unit. Two 100-g subsamples of each forage were also oven-dried for 48 h at 60°C daily in a forced-air oven. Oven DM of ARS and DCC alfalfa silages were 37.3 ± 1.1% and 48.5 ± 1.9%, respectively (mean ± standard deviation). Oven DM of ARS and DCC corn silages were 34.7 ± 1.2% and 37.4 ± 0.5%, respectively (mean ± standard deviation). Dry matter determinations from NIR f were on average 3.5 units higher than the oven DM values. The NIR a DM predictions were on average 1.7 DM units lower than the oven DM values. Additionally, differences among the 3 NIR f probe results were detected (43.1, 40.7, and 41.3% DM, respectively), but all other results were similar between probes. Determinations of DM by the microwave and food dehydrator were also similar with the 60°C, 48-h oven method, whereas the Koster tester was lower than the oven. The handheld NIRS units more accurately predicted DM content of the alfalfa silage but were not as accurate with corn silages when the factory preset calibrations were corrected for bias., (Copyright © 2018 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)

Feed intake is one of the most important components of feed efficiency in dairy systems. However, it is a difficult trait to measure in commercial operations for individual cows. Milk spectrum from mid-infrared spectroscopy has been previously used to predict milk traits, and could be an alternative to predict dry matter intake (DMI). The objectives of this study were (1) to evaluate if milk spectra can improve DMI predictions based only on cow variables; (2) to compare artificial neural network (ANN) and partial least squares (PLS) predictions; and (3) to evaluate if wavelength (WL) selection through Bayesian network (BN) improves prediction quality. Milk samples (n = 1,279) from 308 mid-lactation dairy cows [127 ± 27 d in milk (DIM)] were collected between 2014 and 2016. For each milk spectra time point, DMI (kg/d), body weight (BW, kg), milk yield (MY, kg/d), fat (%), protein (%), lactose (%), and actual DIM were recorded. The DMI was predicted with ANN and PLS using different combinations of explanatory variables. Such combinations, called covariate sets, were as follows: set 1 (MY, BW 0.75 , DIM, and 361 WL); set 2 [MY, BW 0.75 , DIM, and 33 WL (WL selected by BN)]; set 3 (MY, BW 0.75 , DIM, and fat, protein, and lactose concentrations); set 4 (MY, BW 0.75 , DIM, 33 WL, fat, protein, and lactose); set 5 (MY, BW 0.75 , DIM, 33 WL, and visit duration in the feed bunk); set 6 (MY, DIM, and 33 WL); set 7 (MY, BW 0.75 , and DIM); set-WL (included 361 WL); and set-BN (included just 33 selected WL). All models (i.e., each combination of covariate set and fitting approach, ANN or PLS) were validated with an external data set. The use of ANN improved the performance of models 2, 5, 6, and BN. The use of BN combined with ANN yielded the highest accuracy and precision. The addition of individual WL compared with milk components (set 2 vs. set 3) did not improve prediction quality when using PLS. However, when ANN was employed, the model prediction with the inclusion of 33 WL was improved over the model containing only milk components (set 2 vs. set 3; concordance correlation coefficient = 0.80 vs. 0.72; coefficient of determination = 0.67 vs. 0.53; root mean square error of prediction 2.36 vs. 2.81 kg/d). The use of ANN and the inclusion of a behavior parameter, set 5, resulted in the best predictions compared with all other models (coefficient of determination = 0.70, concordance correlation coefficient = 0.83, root mean square error of prediction = 2.15 kg/d). The addition of milk spectra information to models containing cow variables improved the accuracy and precision of DMI predictions in lactating dairy cows when ANN was used. The use of BN to select more informative WL improved the model prediction when combined with cow variables, with further improvement when combined with ANN., (Copyright © 2018 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)

Although cowside testing strategies for diagnosing hyperketonemia (HYK) are available, many are labor intensive and costly, and some lack sufficient accuracy. Predicting milk ketone bodies by Fourier transform infrared spectrometry during routine milk sampling may offer a more practical monitoring strategy. The objectives of this study were to (1) develop linear and logistic regression models using all available test-day milk and performance variables for predicting HYK and (2) compare prediction methods (Fourier transform infrared milk ketone bodies, linear regression models, and logistic regression models) to determine which is the most predictive of HYK. Given the data available, a secondary objective was to evaluate differences in test-day milk and performance variables (continuous measurements) between Holsteins and Jerseys and between cows with or without HYK within breed. Blood samples were collected on the same day as milk sampling from 658 Holstein and 468 Jersey cows between 5 and 20 d in milk (DIM). Diagnosis of HYK was at a serum β-hydroxybutyrate (BHB) concentration ≥1.2 mmol/L. Concentrations of milk BHB and acetone were predicted by Fourier transform infrared spectrometry (Foss Analytical, Hillerød, Denmark). Thresholds of milk BHB and acetone were tested for diagnostic accuracy, and logistic models were built from continuous variables to predict HYK in primiparous and multiparous cows within breed. Linear models were constructed from continuous variables for primiparous and multiparous cows within breed that were 5 to 11 DIM or 12 to 20 DIM. Milk ketone body thresholds diagnosed HYK with 64.0 to 92.9% accuracy in Holsteins and 59.1 to 86.6% accuracy in Jerseys. Logistic models predicted HYK with 82.6 to 97.3% accuracy. Internally cross-validated multiple linear regression models diagnosed HYK of Holstein cows with 97.8% accuracy for primiparous and 83.3% accuracy for multiparous cows. Accuracy of Jersey models was 81.3% in primiparous and 83.4% in multiparous cows. These results suggest that predicting serum BHB from continuous test-day milk and performance variables could serve as a valuable diagnostic tool for monitoring HYK in Holstein and Jersey herds., (Copyright © 2018 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)

The objectives of this study were to investigate the relationship between dry matter intake (DMI) and urinary purine derivative (PD) excretion, to develop equations to predict DMI and to determine the endogenous excretion of PD for beef and dairy cattle using a meta-analytical approach. To develop the models, 62 published studies for both dairy (45 studies) and beef cattle (17 studies) were compiled. Twenty models were tested using DMI (kg/d) and digestible DMI (dDMI, kg/d) as response variables and PD:creatinine (linear term: PD:C, and quadratic term: PD:C 2 ), allantoin:creatinine (linear term: ALLA:C, and quadratic term: ALLA:C 2 ), metabolic body weight (BW 0.75 , kg), milk yield (MY, kg/d), and their combination as explanatory variables for dairy and beef (except for MY) cattle. The models developed to predict DMI for dairy cattle were validated using an independent data set from 2 research trials carried out at the University of Wisconsin (trial 1: n = 45; trial 2: n = 50). A second set of models was developed to estimate the endogenous PD excretion. In all evaluated models, the effect of PD (either as PD:C or ALLA:C) was significant, supporting our hypothesis that PD are in fact correlated with DMI. Despite the BW-independent relationship between PD and DMI, the inclusion of BW 0.75 in the models with PD:C and ALLA:C as predictors slightly decreased the values of root mean square error (RMSE) and Akaike information criterion for the models of DMI. Our models suggest that both DMI and dDMI can be equally well predicted by PD-related variables; however, predicting DMI seems more useful from a practical and experimental standpoint. The inclusion of MY into the dairy models substantially decreased RMSE and Akaike information criterion values, and further increased the precision of the equations. The model including PD:C, BW 0.75 , and MY presented greater concordance correlation coefficient (0.93 and 0.63 for trials 1 and 2, respectively) and lower RMSE of prediction (1.90 and 3.35 kg/d for trials 1 and 2, respectively) when tested in the validation data set, emerging as a potentially useful estimator of nutrient intake in dairy cows. Endogenous PD excretion was estimated by the intercept of the linear regression between DMI (g/kg of BW 0.75 ) and PD excretion (mmol/kg of BW 0.75 ) for beef (0.404 mmol/kg of BW 0.75 ) and dairy cattle (0.651 mmol/kg of BW 0.75 ). Based on the very close agreement between our results for beef cattle and the literature, the linear regression appears to be an adequate method to estimate endogenous PD excretion., (Copyright © 2017 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)

Negative energy balance is an important part of the lactation cycle, and measuring the current energy balance of a cow is useful in both applied and research settings. The objectives of this study were (1) to determine if milk fatty acid (FA) proportions were consistently related to plasma nonesterified fatty acids (NEFA); (2) to determine if an individual cow with a measured milk FA profile is above or below a NEFA concentration, (3) to test the universality of the models developed within the University of Wisconsin and US Dairy Forage Research Center cows. Blood samples were collected on the same day as milk sampling from 105 Holstein cows from 3 studies. Plasma NEFA was quantified and a threshold of 600 µEq/L was applied to classify animals above this concentration as having high NEFA (NEFA high ). Thirty milk FA proportions and 4 milk FA ratios were screened to evaluate their capacity to classify cows with NEFA high according to determined milk FA threshold. In addition, 6 linear regression models were created using individual milk FA proportions and ratios. To evaluate the universality of the linear relationship between milk FA and plasma NEFA found in the internal data set, 90 treatment means from 21 papers published in the literature were compiled to test the model predictions. From the 30 screened milk FA, the odd short-chain fatty acids (C7:0, C9:0, C11:0, and C13:0) had sensitivity slightly greater than the other short-chain fatty acids (83.3, 94.8, 80.0, and 85.9%, respectively). The sensitivities for milk FA C6:0, C8:0, C10:0, and C12:0 were 78.8, 85.3, 80.1, and 83.9%, respectively. The threshold values to detect NEFA high cows for the last group of milk FA were ≤2.0, ≤0.94, ≤1.4, and ≤1.8 g/100 g of FA, respectively. The milk FA C14:0 and C15:0 had sensitivities of 88.7 and 85.0% and a threshold of ≤6.8 and ≤0.53 g/100 g of FA, respectively. The linear regressions using the milk FA ratios C18:1 to C15:0 and C17:0 to C15:0 presented lower root mean square error (RMSE = 191 and 179 µEq/L, respectively) in comparison with individual milk FA proportions (RMSE = 194 µEq/L), C18:1 to even short-medium-chain fatty acid (C4:0-C12:0) ratio (RMSE = 220 µEq/L), and C18:1 to C14:0 (RMSE = 199 µEq/L). Models using milk FA ratios C18:1 to C15:0 and C17:0 to C15:0 had a better fit with the external data set in comparison with the other models. Plasma NEFA can be predicted by linear regression models using milk FA ratios., (Copyright © 2017 American Dairy Science Association. Published by Elsevier Inc. All rights reserved.)