Using predictivity to estimate heritabilities and genetic correlations over time for growth traits in a large...

In this study, we aimed to estimate heritabilities and genetic correlations over time using within- and across-trait predictivity with complete and trimmed references in a large, genotyped population. The data, provided by the American Angus Association, included birth weight (BW), weaning weight (WW), and post-weaning gain (PWG) records from 10,045,726 animals collected from 1972 to 2025, a pedigree of 12,009,755 animals, and genotypes for 1,734,603 of them. We defined nine time slices from 2016 to 2024. The trimmed references included 5 yr of data preceding the year containing the focal group, whereas the full reference used all previous information. Heritabilities remained stable for BW and PWG at around 0.20 and 0.15, respectively. In contrast, heritability for WW increased by 0.05 over time, reaching 0.28 in the full and 0.22 in the trimmed reference. Heritability mirrored changes in predictivity. WW predictivity increased substantially in recent years due to the incorporation of genomic information, which also enhanced the across-trait predictivity whenever WW was involved. The WW-PWG genetic correlation in the early stages of genomic selection was 0.91 in the full and 0.84 in the trimmed reference and decreased slightly over time to 0.84 and 0.80, respectively. In contrast, BW-WW and BW-PWG genetic correlations remained stable around 0.40 and 0.25, respectively. Trimming the reference allows obtaining genetic parameters that reflect the focal group, but it must balance the removal of older generations with retaining enough phenotypes and genotypes to maintain high accuracy. Predictivity enables yearly tracking of genetic parameters, allowing monitoring of possible accelerated effects of genomic selection in rapidly changing traits and correlated traits.