45179

Application of single step GBLUP method under different uncertain paternity scenarios

Rafael Lara Tonussi  ; Rafael Medeiros de Oliveira Silva ; Ana Fabrícia Braga Magalhães  ; Hermenegildo Lucas Justino Chiaia ; Marcos Vinícius Antunes de Lemos ; Mariana Piatto Berton ; Ignacio Aguilar ; Fernando Baldi
Download
Favorite this paper

The aim of this study was to investigate the application of ssGBLUP in different scenarios of uncertain paternity in a simulated beef cattle population. The analyzed population was simulated using the software QMSim 1.00, where ten replicates were performed for two traits and the selected heritability (0.12 for age at first calving and 0.34 for weight at 550 days) was based on estimates of real data and phenotypic variance of 1.0. Scenarios using 0, 25, 50, 75 and 100% of multiple service sires (MS) were studied. The simulated genome had a total length of 2,333 cM, 735,293 markers and 7,000 QTLs, and it assumed that QTLs explained 100% of the genetic variance. Markers and QTLs were randomly distributed over 29 BTA and ranged from 12,931 to 46,495 and from 121 to 438, respectively. All markers were bi-allelic, mimicking SNPs present in the bovine commercial panels. For QTL, the amount of alleles for loci ranged randomly from 2 to 4. Minor allele frequencies (MAF) were assumed equal for markers and QTL alleles. In this method, the inverse of the numerator relationship matrix (A-1) was replaced by H-1, that combines pedigree and genomic information. Accuracies were measured as correlation between the true breeding value (TBV) and the genomic estimated breeding value (GEBV) and the bias by regression coefficient of TBV on GEBV. The accuracy of evaluation and its bias were studied in four animals groups: G1 = all animals; G2 = only bulls; G3 = genotyped animals and G4 = females. The heritability estimates increased as much as the proportion of MS increased for all analyzed traits, which indicated that an increase in the percentage of MS overestimates the additive genetic variance estimation. These results showed that increasing MS decreases the average relationship within family and the number of inbred animals, thus, there is an increase in the genetic variances. The accuracy of evaluation ranged from 0.557 to 0.869 and from 0.656 to 0.971 for AFC and W550, respectively, with more accurate prediction in the G2 and G4. An increase in the proportion of RM in the population decreased the accuracies in all groups in both studied traits. The regression coefficient of TBV on GEBV of studied groups were near to 1, which indicated that the predictions were not biased.