Are Past Generations Contributing to Evaluations on Young Genotyped Animals?

Abstract Text: Datasets of US and Israeli Holsteins, and pigs from PIC were used to evaluate the impact of different number of generations on ability to predict GEBV of young genotyped animals. The inclusion of only two generations of ancestors (A2) or all ancestors (Af) was also evaluated. A total of 34,506 US and 1,305 Israeli Holsteins bulls, and 5,236 pigs were genotyped. The evaluations were computed by traditional BLUP and single-step GBLUP, with respective computing performance recorded. For the two Holstein datasets, coefficients of determination and regression of deregressed evaluations from a full dataset with records up to 2011 on EBV or GEBV from the reduced dataset (up to 2006 for Israeli and 2007 for US) and truncations were computed. The thresholds for old data deletion were based on generation intervals of 5 years. For the PIC dataset, correlations between corrected phenotypes and EBV or GEBV were used to evaluate the predictive ability on young animals born in 2010 and 2011. The reduced dataset contained data up to 2009 and the thresholds were based on generation interval of 3 years. The number of generations that could be deleted without reduction in accuracy was dependent on data structure and trait. For US Holsteins, removing 3 and 4 generations of data did not reduce accuracy of evaluations for final score in Af and A2 scenarios, respectively. For Israeli Holsteins, the accuracies for milk, fat, and protein yields were the highest when only phenotypes recorded on year ≥ 2000 and full pedigrees were included. Of the 135 Israeli validation bulls with genotypes and daughter records only in the complete dataset, 38 and 97 were sons of Israeli and foreign bulls, respectively. While more phenotypic data increased the prediction accuracy for sons of Israeli bulls, the reverse was true for sons of foreign bulls. For PIC dataset, removing data up to five generations did not erode predictive ability for genotyped animals for litter size and number of stillborn. Given the data used in this study, truncating old data does not decrease the accuracy on young genotyped animals, while reducing computation requirements and helping to find problems due to population structure. For populations that include local and imported animals, the truncation may be beneficial for one group of animals and detrimental to another.

Keywords: ssGBLUP, pedigree depth, genomic selection