Skip to main content
Figure 2 | Genetics Selection Evolution

Figure 2

From: Accuracy of genomic predictions in Bos indicus (Nellore) cattle

Figure 2

Empirical accuracies of genomic predictions for 15 traits ** of Bos indicus (Nellore) cattle for different validation strategies * . *forward = training set composed of bulls with highly accurate EBV in 2007 and testing set composed of the remaining bulls (with accurate EBV in 2011); RAND = 5-fold cross-validation (CV), splitting animals randomly into groups of similar size; DIST = 5-fold cross-validation, based on k-means clustering of animals based on their genomic distance (i.e. minimizing inter-groups relationships); empirical accuracies were calculated as the Pearson’s correlation between DGV (obtained with GBLUP20) and deregressed EBV (dEBV) in 2011 for the testing set, divided by the average accuracy of dEBV in the testing set; for the cross-validation strategies (RAND and DIST), the bars and errors bars represent the estimates of means and standard errors obtained in 5-fold CV, respectively; **WG = weight gain from birth to weaning (about 205 days of age); Cw, Pw, Mw, Nw = visual scores taken at weaning for carcass conformation, finishing precocity, muscling and navel, respectively; PWG = weight gain from weaning to yearling (about 550 days of age); Cy, Py, My, Ny = visual scores taken at yearling for carcass conformation, finishing precocity, muscling and navel, respectively; SCaw = scrotal circumference adjusted for age and weight; BW = birth weight; GL = gestation length; WI = weaning index, composed of traits evaluated at weaning; FI = final index, composed by traits evaluated at weaning and yearling (FI) [See Additional file 1 for more details].

Back to article page