Disentangling data dependency using cross-validation strategies to evaluate prediction quality of cattle grazing activities using machine learning algorithms and wearable sensor data

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.     

Suckling behavior,performance, and Kleiber index in calves of different genetic groups under shaded and unshaded environments

Tropical Animal Health and Production - According to previous studies, lamb mortality is high in the Ethiopian highlands. The present study aims to evaluate the execution of preventive sheep herd...     

Morphological and molecular characterization of the sudden-death syndrome pathogen of soybean in Brazil

Brazilian Fusarium isolates causing soybean sudden death syndrome (SDS) were characterized by comparing them with other Fusarium isolates associated with soybean root rot, as well as F. solani f.sp. glycines isolates associated with the disease in the USA, using molecular (mitochondrial and nuclear rDNA), morphological, cultural and pathogenic characteristics. On the basis of pathogenicity data, and restriction fragment length polymorphism and sequence analysis of the rDNA internal transcribed spacer (ITS) regions, isolates formed a group distinct from nonSDS F. solani isolates, as well as other Fusarium species. ITS sequence analysis also revealed that Brazilian isolates were distinct from the majority of SDS pathogens from the USA ( Fusarium virguliforme ) and conformed to Fusarium tucumaniae .     

High frequency of porcine norovirus infection in finisher units of Brazilian pig-production systems

Tropical Animal Health and Production - Norovirus (NoV) is a member of the Caliciviridae family and is considered an emerging human enteric pathogen. NoVs are detected in farm animals such as...     

Descriptive analysis and spatial epidemiology of porcine reproductive and respiratory syndrome (PRRS) for swine sites participating in area regional control and elimination programs from 3 regions of Ontario

The objectives of this study were to describe demographics, basic biosecurity practices, ownership structure, and prevalence of porcine reproductive and respiratory syndrome (PRRS) in swine sites located in 3 regions in Ontario, and investigate the presence of spatial clustering and clusters of PRRS positive sites in the 3 regions. A total of 370 swine sites were enrolled in Area Regional Control and Elimination projects in Niagara, Watford, and Perth from 2010 to 2013. Demographics, biosecurity, and site ownership data were collected using a standardized questionnaire and site locations were obtained from an industry organization. Status was assigned on the basis of available diagnostic tests and/or assessment by site veterinarians. Spatial dependence was investigated using the D-function, the spatial scan statistic test and the spatial relative risk method. Results showed that the use of strict all-in all-out (AIAO) pig flow and shower before entry are uncommon biosecurity practices in swine sites, but a larger proportion of sites reported having a Danish entry. The prevalence of PRRS in the 3 regions ranged from 17% to 48% and localized high and low risk clusters were detected. Sites enrolled in the PRRS control projects were characterized by membership in multiple and overlapping ownership structures and networks, which complicates the way the results of monitoring and disease management measures are communicated to the target population.     

Evaluation of growth and fatty acid profile of Nile tilapia (Oreochromis niloticus) fed with Schizochytrium sp.

The objective of this study was to evaluate growth, body chemical composition and lipid profile of Nile tilapia juvenile fed with Schizochytrium sp. Two hundred and forty Nile tilapia (Oreochromis niloticus) juvenile (1.33 ± 0.11 g) were distributed in 20 aquariums, at the density of 12 fish per aquarium. The juvenile were fed with five levels of Schizochytrium sp.: Control (0 g/kg of Schizochytrium sp. in feed); AS10 (10 g/kg of Schizochytrium sp. in feed); AS20 (20 g/kg of Schizochytrium sp. in feed); AS30 (30 g/kg of Schizochytrium sp. in feed) and AS40 (40 g/kg of Schizochytrium sp. in feed). The inclusion of Schizochytrium sp. increase the body weight, weight gain, final biomass and biomass gain of tilapia juvenile. The body crude protein of tilapia was increased after addition of microalgae. Juvenile lipid profile also was influenced when fed with 40 g of Schizochytrium sp. per kg and have it levels of omega‐3 and docosahexaenoic acid (DHA) increased, and omega 6:3 ratio decreased. Tilapia juvenile fed with Schizochytrium sp. per kg have better growth, omega‐3 and DHA levels rising in the body, being a considerable source of fatty acids for human nutrition. Levels above 20 g of Schizochytrium sp. per kg in the feed favour the increase of protein in Nile tilapia body.     

Sampling of Sugarcane Leaves in Field Experiments to Determine the Activity of Nitrate Reductase

Nitrogen (N) assimilation depends on nitrate reductase activity (NRA). Leaf sampling methods in field experimentation with sugarcane and analytical procedures are essential for the determination of the maximum activity (MA) of the enzyme. Thus, in five tests, it was possible to determine which period of the crop cycle should be used for leaf collection, the most adequate leaf for collection, the ideal number of leaves to be collected, the period of the day in which the enzyme has MA and which analytical procedures can be optimized. The evaluation of NRA in sugarcane should occur between 95 and 100 days after planting (DAP) and can be performed in any green leaf. In field experiments, it is recommended the collection of 15 leaves, which must occur between 12:00 and 12:30 h, with later incubation of the leaf tissues for 430 min, without vacuum application.     

Characterization of Isoforms of the Lectin Isolated from the Red Algae Bryothamnion seaforthii and Its Pro-Healing Effect

Lectins are a structurally heterogeneous group of proteins that have specific binding sites for carbohydrates and glycoconjugates. Because of their biotechnological potential, lectins are widely used in biomedical research. The present study aimed to evaluate the healing potential of the lectin isolated from the marine red alga Bryothamnion seaforthii (BSL). The lectin was purified using ion exchange chromatography with DEAE cellulose and characterized using tandem mass spectrometry. For healing tests, skin wounds were induced in the dorsal thoracic region of mice. These animals were randomly divided into three groups and subjected to topical treatment for 12 days with BSL, bovine serum albumin and 150 mM NaCl. To evaluate the potential of each treatment, the animals were anesthetized and sacrificed on days 2, 7 and 12, respectively. The parameters evaluated included the wound area, the proportion of wound closure and the histological diagnosis. The wound closure was more effective with BSL (Postoperative Day 7 and 12) than controls. The luminal epithelium was completely restructured; the presence of collagen in the dermis and the strongly active presence of young skin annexes demonstrate the potential of treatment with BSL compared with controls. Our findings suggest that BSL has pro-healing properties and can be a potential medical process in the treatment of acute wounds.