The aim of this study was to investigate the effects of two feed supplements on rumen bacterial communities of heifers fed a high grain diet. Six Holstein-Friesian heifers received one of the following dietary treatments according to a Latin square design: no supplement (control, C), 60 g/day of fumarate-malate (organic acid, O) and 100 g/day of polyphenol-essential oil (P). Rumen fluid was analyzed to assess the microbial population using Illumina sequencing and quantitative real time PCR. The P treatment had the highest number of observed species (P < 0.10), Chao1 index (P < 0.05), abundance based coverage estimated (ACE) (P < 0.05), and Fisher’s alpha diversity (P < 0.10). The O treatment had intermediate values between C and P treatments with the exception of the Chao1 index. The PCoA with unweighted Unifrac distance showed a separation among dietary treatments (P = 0.09), above all between the C and P (P = 0.05). The O and P treatments showed a significant increase of the family Christenenellaceae and a decline of Prevotella brevis compared to C. Additionally, the P treatment enhanced the abundance of many taxa belonging to Bacteroidetes, Firmicutes and Tenericutes phyla due to a potential antimicrobial activity of flavonoids that increased competition among bacteria. Organic acid and polyphenols significantly modified rumen bacterial populations during high-grain feeding in dairy heifers. In particular the polyphenol treatment increased the richness and diversity of rumen microbiota, which are usually high in conditions of physiological rumen pH and rumen function.

The aim of the present study was to develop a model to identify posture and behavior from data collected by a triaxial accelerometer located on the left flank of dairy cows and evaluate its accuracy and precision. Twelve Italian Red-and-White lactating cows were equipped with an accelerometer and observed on average for 136 ± 29 min per cow by two trained operators as a reference. The acceleration data were grouped in time windows of 8 s overlapping by 33.0%, for a total of 35,133 rows. For each row, 32 different features were extracted and used by machine learning algorithms for the classification of posture and behavior. To build up a predictive model, the dataset was split in training and testing datasets, characterized by 75.0 and 25.0% of the observations, respectively. Four algorithms were tested: Random Forest, K Nearest Neighbors, Extreme Boosting Algorithm (XGB), and Support Vector Machine. The XGB model showed the best accuracy (0.99) and Cohen’s kappa (0.99) in predicting posture, whereas the Random Forest model had the highest overall accuracy in predicting behaviors (0.76), showing a balanced accuracy from 0.96 for resting to 0.77 for moving. Overall, very accurate detection of the posture and resting behavior were achieved.

The aim of this study was to measure the level of activity and rumination in young bulls and to assess whether these data can be used as indicators of health status and average daily weight gain (ADG). Two groups of animals (period 1: n=108 animals; period 2: n=106 animals) were fitted with sensors to measure daily activity and rumination, were weighed on arrival and at the end of the trial (70 days) and were checked twice daily to verify their health condition. Any clinical signs and therapies were recorded. The dishomogeneity index of rumination (DR), and the daily dishomogeneity indices of activity (DDA) and rumination (DDR), were calculated. Bulls had an ADG of 1.42±0.38kg/day and showed an average duration of daily rumination of 404±63min and an average activity of 474±46 bits, respectively. Animals characterised by low ADG had lower values of minimum daily rumination (P=0.01) and DDA (P<0.001), and a greater rumination range (P=0.007) and DR (P=0.003). Bovine respiratory disease and lameness were detected 31 and five times, respectively; among affected animals, the average daily activity, rumination and DDA were lower (P<0.05) at 3-6days before the onset of visible clinical signs, whereas DDR increased compared to the values when individuals were apparently healthy. The use of individual sensors appears promising for the early diagnosis of disease in beef cattle and for improving herd management.

The accurate detection of behavioural changes represents a promising method of detecting the early onset of disease in dairy cows. This study assessed the performance of deep learning (DL) in classifying dairy cows’ behaviour from accelerometry data acquired by single sensors on the cows’ left flanks and compared the results with those obtained through classical machine learning (ML) from the same raw data. Twelve cows with a tri-axial accelerometer were observed for 136 ± 29 min each to detect five main behaviours: standing still, moving, feeding, ruminating and resting. For each 8 s time interval, 15 metrics were calculated, obtaining a dataset of 211,720 observation units and 15 columns. The entire dataset was randomly split into training (80%) and testing (20%) datasets. The DL accuracy, precision and sensitivity/recall were calculated and compared with the performance of classical ML models. The best predictive model was an 8-layer convolutional neural network (CNN) with an overall accuracy and F1 score equal to 0.96. The precision, sensitivity/recall and F1 score of single behaviours had the following ranges: 0.93–0.99. The CNN outperformed all the classical ML algorithms. The CNN used to monitor the cows’ conditions showed an overall high performance in successfully predicting multiple behaviours using a single accelerometer.

Background: Ruminal acidosis is responsible for the onset of different pathologies in dairy and feedlot cattle, but there are major difficulties in the diagnosis. This study modelled the data obtained from various blood variables to identify those that could indicate the severity of ruminal acidosis. Six heifers were fed three experimental rations throughout three periods. The diets were characterised by different starch levels: high starch (HS), medium starch (MS) and low starch, as the control diet (CT). Ruminal pH values were continuously measured using wireless sensors and compared with pH measurements obtained by rumenocentesis. Blood samples were analysed for complete blood count, biochemical profile, venous blood gas, blood lipopolysaccharide (LPS) and LPS-binding proteins (LBP). Results: The regression coefficient comparing the ruminal pH values, obtained using the two methods, was 0.56 (P = 0.040). Feeding the CT, MS and HS led to differences in the time spent below the 5.8, 5.5 and 5.0 pH thresholds and in several variables, including dry matter intake (7.7 vs. 6.9 vs. 5.1 kg/d; P = 0.002), ruminal nadir pH (5.69 vs. 5.47 vs. 5.44; P = 0.042), mean ruminal pH (6.50 vs. 6.34 vs. 6.31; P = 0.012), haemoglobin level (11.1 vs. 10.9 vs. 11.4 g/dL; P = 0.010), platelet count (506 vs. 481 vs. 601; P = 0.008), HCO₃ ^- (31.8 vs. 31.3 vs. 30.6 mmol/L; P = 0.071) and LBP (5.9 vs. 9.5 vs. 10.5 μg/mL; P < 0.001). A canonical discriminant analysis (CDA) was used to classify the animals into four ruminal pH classes (normal, risk of acidosis, subacute ruminal acidosis and acute ruminal acidosis) using haemoglobin, mean platelet volume, β-hydroxybutyrate, glucose and reduced haemoglobin. Conclusions: Although additional studies are necessary to confirm the reliability of these discriminant functions, the use of plasma variables in a multifactorial model appeared to be useful for the evaluation of ruminal acidosis severity.