Diagnosis of Bovine Respiratory Disease in feedlot cattle using blood 1H NMR metabolomics

C. Blakebrough-Hall,A. Dona,M. J. D’occhio,J. McMeniman,L. A González
DOI: https://doi.org/10.1038/s41598-019-56809-w
IF: 4.6
2020-01-10
Scientific Reports
Abstract:Abstract Current diagnosis methods for Bovine Respiratory Disease (BRD) in feedlots have a low diagnostic accuracy. The current study aimed to search for blood biomarkers of BRD using 1 H NMR metabolomics and determine their accuracy in diagnosing BRD. Animals with visual signs of BRD (n = 149) and visually healthy (non-BRD; n = 148) were sampled for blood metabolomics analysis. Lung lesions indicative of BRD were scored at slaughter. Non-targeted 1 H NMR metabolomics was used to develop predictive algorithms for disease classification using classification and regression trees. In the absence of a gold standard for BRD diagnosis, six reference diagnosis methods were used to define an animal as BRD or non-BRD. Sensitivity (Se) and specificity (Sp) were used to estimate diagnostic accuracy (Acc). Blood metabolomics demonstrated a high accuracy at diagnosing BRD when using visual signs of BRD (Acc = 0.85), however was less accurate at diagnosing BRD using rectal temperature (Acc = 0.65), lung auscultation score (Acc = 0.61) and lung lesions at slaughter as reference diagnosis methods (Acc = 0.71). Phenylalanine, lactate, hydroxybutyrate, tyrosine, citrate and leucine were identified as metabolites of importance in classifying animals as BRD or non-BRD. The blood metabolome classified BRD and non-BRD animals with high accuracy and shows potential for use as a BRD diagnosis tool.
multidisciplinary sciences
What problem does this paper attempt to address?