Novel approaches to assess the quality of fertility data stored in dairy herd management software.

Scientific journals and popular press magazines are littered with articles in which the authors use data from dairy herd management software. Almost none of such papers include data cleaning and data quality assessment in their study design despite this being a very critical step during data mining. This paper presents 2 novel data cleaning methods that permit identification of animals with good and bad data quality. The first method is a deterministic or rule-based data cleaning method. Reproduction and mutation or life-changing events such as birth and death were converted to a symbolic (alphabetical letter) representation and split into triplets (3-letter code). The triplets were manually labeled as physiologically correct, suspicious, or impossible. The deterministic data cleaning method was applied to assess the quality of data stored in dairy herd management from 26 farms enrolled in the herd health management program from the Faculty of Veterinary Medicine Ghent University, Belgium. In total, 150,443 triplets were created, 65.4% were labeled as correct, 17.4% as suspicious, and 17.2% as impossible. The second method, a probabilistic method, uses a machine learning algorithm (random forests) to predict the correctness of fertility and mutation events in an early stage of data cleaning. The prediction accuracy of the random forests algorithm was compared with a classical linear statistical method (penalized logistic regression), outperforming the latter substantially, with a superior receiver operating characteristic curve and a higher accuracy (89 vs. 72%). From those results, we conclude that the triplet method can be used to assess the quality of reproduction data stored in dairy herd management software and that a machine learning technique such as random forests is capable of predicting the correctness of fertility data.

Practices to optimise gastrointestinal nematode control on sheep, goat and cattle farms in Europe using targeted (selective) treatments.

Due to the development of anthelmintic resistance, there have been calls for more sustainable nematode control practices. Two important concepts were introduced to study and promote the sustainable use of anthelmintics: targeted treatments (TT), where the whole flock/herd is treated based on knowledge of the risk, or parameters that quantify the severity of infection; and targeted selective treatments (TST), where only individual animals within the grazing group are treated. The aim of the TT and TST approaches is to effectively control nematode-induced production impacts while preserving anthelmintic efficacy by maintaining a pool of untreated parasites in refugia. Here, we provide an overview of recent studies that assess the use of TT/TST against gastrointestinal nematodes in ruminants and investigate the economic consequences, feasibility and knowledge gaps associated with TST. We conclude that TT/TST approaches are ready to be used and provide practical benefits today. However, a major shift in mentality will be required to make these approaches common practice in parasite control.

Milk fat saturation and reproductive performance in dairy cattle.

Unsaturated fatty acids (UFA) cannot be synthesized by mammalian cells due to a lack of desaturase enzymes. Combined with their limited supply to the small intestines, UFA have been proposed as nutraceuticals to ameliorate dairy cow fertility. However, field studies based on a large number of animals are lacking on this subject. Therefore the aim of the present study was to analyze a large dataset containing individual cow fertility records from dairy herds and link fertility key-performance-indicators like conception rate to first insemination (CRFI), days in milk to first insemination (DIMFI) and days in milk to conception (DIMCONC), to the level of UFA in bulk tank samples, the latter being a proxy for the dietary fatty acid profile on these herds. Within the two year study period, information from 15,055 lactations and 35,433 bulk tank milk samples was collected on 90 herds. The multilevel logistic regression model used, revealed a decreased CRFI on herds with a higher bulk tank UFA level. The decrease in CRFI was larger for higher producing herds. Increased bulk tank UFA was furthermore associated with higher DIMFI which, together with the lower CRFI, subsequently increased DIMCONC. Interestingly, higher variability in UFA, expressed by an increased coefficient of variation, was associated with an increased CRFI and decreased DIMFI and DIMCONC. In conclusion, the present study demonstrates that increasing the UFA content of milk should not be a goal as such when supplementing UFA to dairy cows as higher bulk tank UFA are associated with worsened fertility results.

On-farm evaluation of the effect of metabolic diseases on the shape of the lactation curve in dairy cows through the MilkBot lactation model.

The effects of metabolic diseases (MD) occurring during the transition period on milk production of dairy cows have been evaluated in many different ways, often with conflicting conclusions. The present study used a fitted lactation model to analyze specific aspects of lactation curve shape and magnitude in cows that avoided culling or death in the first 120 d in milk (DIM). Production and health records of 1,946 lactations in a 1-yr follow-up study design were collected from a transition management facility in Germany to evaluate both short- and long-term effects of MD on milk production. Milk production data were fitted with the nonlinear MilkBot lactation model, and health records were used to classify cows as healthy (H), affected by one MD (MD), or by multiple MD (MD+). The final data set contained 1,071 H, 348 MD, and 136 MD+ cows, with distinct incidences of 3.7% twinning, 4.8% milk fever, 3.6% retained placenta, 15.4% metritis, 8.3% ketosis, 2.0% displaced abomasum, and 3.7% mastitis in the first 30 DIM. The model containing all healthy and diseased cows showed that lactations classified as H had milk production that increased faster (lower ramp) and also declined faster (lower persistence) compared with cows that encountered one or more metabolic problems. The level of production (scale) was only lowered in MD+ cows compared with H and MD cows. Although the shape of the lactation curve changed when cows encounter uncomplicated (single) MD or complicated MD (more than one MD), the slower increase to a lower peak seemed to be compensated for by greater persistency, resulting in the overall 305-d milk production only being lowered in MD+ cows. In the individual disease models, specific changes in the shape of the lactation curve were found for all MD except twinning. Milk fever, retained placenta, ketosis, and mastitis mainly affected the lactation curve when accompanied by another MD, whereas metritis and displaced abomasum affected the lactation curve equally with or without another MD. Overall, 305-d milk production was decreased in complicated metritis (10,603 ± 50 kg vs. 10,114 ± 172 kg). Although care should be taken in generalizing conclusions from a highly specialized transition management facility, the current study demonstrated that lactation curve analysis may contribute substantially to the evaluation of both short- and long-term effects of metabolic diseases on milk production by detecting changes in the distribution of production that are not apparent when only totals are analyzed.