Structure discovery in Bayesian networks: an analytical tool for analysing complex animal health data

Analysing animal health data can be a complex task as the health status of individuals or groups of animals, might depend on many inter-related variables. The objective is to differentiate variables that are directly associated with health status and therefore promising targets for intervention, from variables that are indirectly associated with health status and can therefore at best only affect this indirectly through association with other variables. Bayesian network (BN) modelling is a machine learning technique for empirically identifying associations in complex and high dimensional data, so-called "structure discovery". An introduction to structure discovery using BN modelling is presented, comprising the key assumptions required by the methodology, along with a discussion of advantages and limitations. To demonstrate the various steps required to apply BN structure discovery to animal health data, illustrative analyses of data collected during a previously published study concerned with exposure to bovine viral diarrhoea virus in beef cow-calf herds in Scotland are presented.     

Hepatic dearterialization in the dog

Hepatic dearterialization was performed in 9 adult dogs. Evaluation of the technique was based on clinical and laboratory assessment of physiologic and pathologic changes related to the procedure. All dogs had hematologic and histopathologic evidence of transient hepatic ischemia that was reflected by minimal post-operative morbidity.     

Accidental sulfur poisoning in a group of holstein heifers

Fourteen animals died or were euthanized after toxic levels of elemental sulfur were accidentally fed to a group of 120 Holstein heifers. Dehydration, rumen stasis, tachycardia, and diarrhea were seen along with metabolic acidosis, hypokalemia, and hypochloremia. The majority of deaths occurred from 3 to 10 days after the sulfur was fed to the heifers. Postmortem examination showed rumenitis, acute alveolitis, and renal tubular necrosis. The toxicity of ingested sulfur was attributed to the conversion of sulfur to hydrogen sulfide in the rumen.     

Bartonella henselae infection in a dog with recalcitrant ineffective erythropoiesis

Ineffective erythropoiesis was diagnosed in an 8‐year‐old male castrated Labrador Retriever. Despite treatment with immunosuppressive therapy for suspected immune‐mediated erythrocyte maturation arrest, resolution of the nonregenerative anemia was not achieved. Following documentation of Bartonella henselae bacteremia by Bartonella alpha proteobacteria growth medium (BAPGM) enrichment blood culture, immunosuppressive therapy was discontinued, and the anemia resolved following prolonged antibiotic therapy. Bartonella immunofluorescent antibody testing was negative, whereas B henselae western blot was consistently positive. The contribution of B henselae bacteremia to ineffective erythropoiesis remains unknown; however, the potential role of B henselae in the pathophysiology of bone marrow dyscrasias warrants additional investigation.     

Impaired littoral energy pathways cause a shift to pelagic resources by fish in recovering lake food webs

Fish communities may increase in biomass and productivity due to energy subsidies from the littoral invertebrate community. In lakes recovering from acidification and metal contamination, such as those in Sudbury, Ontario, Canada, impaired benthic invertebrate communities (i.e., low diversity with higher abundance of small‐bodied taxa) allowed a critical test of the role of these littoral pathways on fish diet. We compared fish abundance, diversity, diet and biomass in eight recovering and eight reference lakes and related availability of the main littoral and pelagic invertebrate groups to fish diet regime using stable isotope analysis. A Bayesian mixing model (MixSIR) was used to estimate diet likelihood, and convex hull analysis was used to estimate trophic niche space of fish communities. Fish biomass did not differ between impaired and reference lakes despite substantial differences in potential diet. Fish depended strongly on littoral benthos in the reference lakes but consumed more pelagic food in the impaired lakes. The trophic niche of the focal, most common fish species (i.e., yellow perch, smallmouth bass, pumpkinseed and brown bullhead) was larger in the impaired lakes. We attributed these differences to low diversity at the highest trophic levels of fish communities in the impaired lakes as well as to depauperate benthic invertebrate communities. In contrast to the food webs of most temperate lakes, fish in impaired lakes preyed less on littoral invertebrates yet still managed to maintain a reference lake level of biomass standing crop by relying more on pelagic resources – macro zooplankton such as Chaoborus.