Random forests prediction of blood metabolic clusters by milk metabolites and enzymes in dairy cows

Research output: Contribution to book/anthology/report/proceedingConference abstract in proceedingsResearchpeer-review

Milk biomarkers as predictors of physiological imbalances and subclinical or clinical diseases in dairy cows may be used in management and selection. The objective was to use milk metabolites and enzymes to predict three clusters differentiating the metabolic status of cows and defined by clustering of plasma hormone Insulin-like Growth Factor-1 (IGF-1) and three plasma metabolites: Glucose, beta-hydroxybutyrate (BHB), and non-esterified fatty acids (NEFA). Data between calving and 50 days in milk (DIM) were obtained from 209 Holstein Friesian cows in six research herds: AU (Denmark), UCD (Ireland) AFBI (UK), CRA-W (Belgium), FBN (Germany), CREA (Italy). Some diets were designed to challenge the cow and provoke production diseases but there were too few incidences to establish predictors of clinical and subclinical diseases. Instead, k-means clustering was performed over scaled residuals from linear mixed effects models for each of IGF-1, Glucose, log10(BHB) and log10(NEFA) measured around 14 and 35 DIM. Each milk metabolite (free glucose, glucose-6-phosphate, BHB, isocitrate, uric acid, urea) and enzyme (LDH, NAGase) from 2 milkings/week was averaged within week and matched with blood sampling. Random forests (RF) predictions with milk markers and parity (1, 2 and 3+) was evaluated by leave-one-cow-out (internal) cross-validation. Overall accuracy: 0.52 (95% CI: 0.45-0.59) on DIM14 and 0.44 (0.37-0.51) on DIM35. Sensitivity for each cluster: (0.77, 0.87, 0.61) on DIM14 and (0.65, 0.71, 0.79) on DIM35. Specificity: (0.54, 0.34, 0.61) and (0.46, 0.51, 0.32). We obtained predictions of metabolic clusters from milk biomarkers of fair accuracy. Sample sizes may constraint both estimation of clusters, number of classes, and RF training. External validation in separate herds is warranted. Still, we see potential in using milk biomarkers to support management of dairy cows.
Original languageEnglish
Title of host publicationBook of Abstracts of the 68th Annual Meeting of the European Federation of Animal Science
Place of publicationThe Netherlands
PublisherWageningen Academic Publishers
Publication year2017
ISBN (print)978-90-8686-312-9
ISBN (Electronic)978-90-8686-859-9
Publication statusPublished - 2017
Event68th Annual Meeting of the European Federation of Animal Science : EAAP - Nordea Concert Hall, Tallinn, Estonia
Duration: 28 Aug 20171 Sep 2017


Conference68th Annual Meeting of the European Federation of Animal Science
LocationNordea Concert Hall
SeriesBook of Abstracts of the Annual EAAP Meeting

See relations at Aarhus University Citationformats


ID: 120130784