PSIV-27 Genomic prediction of novel profitability traits in feedlot under different approaches in Nelore cattle.

Abstract Several economic traits used in genetic improvement programs have helped increase production, but most are evaluated in pre- and post-weaning periods. After the yearling period, few traits are measured directly in the animal, with some only being indicators in the finishing phase, leaving a gap in information and responses to be explored from a bioeconomic point of view. The incorporation of new characteristics in the finishing phase to identify more efficient animals is of great importance since this phase represents the peak of the animal’s productive potential. Therefore, the aim was to assess the accuracy, bias, and dispersion of genomic predictions accumulated profitability (AFP) and profit per kilogram of liveweight gain (PFT) in Nelore cattle using different prediction approaches. The data set consisted of 3,969 phenotypic records for each trait. The pedigree harbored information from 38,930 animals born between 1998 and 2016, 2,691 sires, and 19,884 dams. A total of 2,449 animals were genotyped with the Clarifide® Nelore 3.0 SNP panel. Nine models for genomic prediction were evaluated: a linear animal model was applied to estimate the genetic parameters and to perform the genomic single-trait best linear unbiased prediction (ST_ss - default), bi-trait ssGBLUP (TT_CAR, TT_W450, and TT_DMI), and multi-trait ssGBLUP (MT_ss), and finally, two models using the weighted linear (ST_sswl1 and ST_sswl2) and nonlinear (ST_sswnl1 and ST_sswnl2) single-step genomic approach (WssGBLUP) were implemented to predict genomic breeding values (GEBV). The ability to predict future performance was calculated as the correlation between GEBV and adjusted phenotypes. The average prediction accuracy of the GEBV of the models ranged from 0.345 to 0.665 for PFT and from 0.425 to 0.603 for APF. The predictive capability of the MT_ss model (0.665) was significantly higher than that of the other models for PFT, except for the TT_CAR model (0.604), which also showed improvements in predictive capacity. For APF, the MT_ss (0.561) and TT_W450 (0.556) models demonstrated improvements in genomic prediction accuracy compared to the other models. In general, the single trait ssGBLUP (ST_ss – default) models and the nonlinear weighting did not increase the accuracy of predictions for both traits. For the phenotypic prediction ability of PFT, the linear WssGBLUP models ST_sswl1 (0.65) and ST_sswl2 (0.70), TT_W450 W450 (0.64), and ssGBLUP-M (0.66) demonstrated the highest prediction abilities. Similar results were observed for the phenotypic prediction ability of AFP for both models. However, the linear WssGBLUP model ST_sswl1 (0.84) and ST_sswl2 (0.94) provided higher prediction compared to the bi-trait and multi-trait models. The results indicate that the multi-trait model achieved better predictive ability for the new traits of PFT and APF. Multi-trait genomic selection may provide greater genetic gains than other models for these new economically important traits in breeding programs.