237 Optimizing computing time of genetic evaluations through data truncation in Angus cattle.

Abstract Single-step GBLUP enables the evaluation of any population size, providing accurate genomic breeding values (GEBV). However, in large genomic models, computing time is costly, raising the question of whether using the full dataset is truly beneficial or if fewer data can achieve similar results while reducing computational costs. The objective of this study was to reduce the computing time of the American Angus Association growth model through data truncation. The traits analysed were birth weight (BW), weaning weight (WW), and post-weaning gain (PWG). The initial dataset included 12,802,165 animals with records and 1,570,859 genotyped. Phenotypic truncation removed records for animals born before 1985 (DT85), 1995 (DT95), 2005 (DT05), or 2015 (DT15), with DT0 retaining all records. Genotypic truncation excluded animals without records (GT1), while GT0 applied no truncation. Indirect predictions (IP) were used to estimate genetic merit for animals excluded from the main evaluation. We validated the GEBV using the LR and predictive ability methods. The LR prediction accuracy across the scenarios was in the range of 0.62-0.63 (BW), 0.74-0.77 (WW), and 0.72-0.74 (PWG). Predictivity for DT0, DT85, DT95, and DT05 was 0.51 for BW, 0.47 for WW, and 0.35 for PWG. The values for DT15 were 0.01 lower than these. Correlations GEBV-IP were ☐ 0.99 for DT85, DT95, and DT05. The results showed that moderate phenotypic and genotypic data truncation (DT05-GT1) was suitable, reducing computing time by 66% without compromising the model’s ability to predict future phenotypes. These results indicate that old records have limited impact on the genetic predictions of young animals, and that young genotyped animals without phenotypes contribute little to the prediction of their relatives. Indirect predictions offered a fast and reliable strategy to evaluate non-informative animals excluded in the GT1 scenario. Data truncation can maintain prediction accuracy and preserve dispersion, particularly when working with robust phenotypic and genotypic datasets, non-selective genotyping, moderate to high heritability traits, and a pedigree limited to three to four generations.