Reproduction

Genetic evaluation of gestation length in Holstein cattle: comparison between fetal and dam genetic effects.

---

Abstract

Gestation length (GL) is linked to calving performance, as shorter gestations are related to low birth weight and reduced neonatal viability, whereas longer gestations are associated with heavier calves and increased calving difficulty, highlighting the importance of including GL in Holstein cattle breeding programs. This study aimed to estimate variance components and genetic parameters using different models that considered GL as a cow or fetus phenotype, and to evaluate the contribution of genomic information through the single-step genomic BLUP (ssGBLUP) approach in Holstein cattle. The database, provided by the Associação Brasileira de Criadores de Bovinos da Raça Holandesa (ABCBRH) and Embrapa Dairy Cattle, included a total of 339,510 GL records and the pedigree included 638,176 animals. A total of 16 models were considered and divided into two sets of eight, according to whether GL was defined as a cow or fetus phenotype. The models differed in the inclusion of additive genetic effects (fetus, cow, paternal, and maternal), service sire effects, permanent environment effect of the cow, service sire, and sire. Model comparison was based on the Deviance Information Criterion (DIC), estimates of genetic parameters, and predictive capacity with complete and reduced data. Spearman's rank correlation coefficient between the predicted genetic values (EBVs) was used to evaluate the re-ranking of animals across models. The contribution of genomic information for the GL was also evaluated. Genotypes were imputed to the 50Kv2 SNP chip using FImpute 3.0. Quality control criteria included a SNP call rate of at least 95%, minor allele frequency (MAF) of 0.02, and Hardy-Weinberg equilibrium threshold of χ² ≤ 10⁻⁶. Analyses were conducted using the BLUPF90 software family, and SNP effects were estimated using postGSf90 with Bonferroni correction. Heritability estimates ranged from 0.20 to 0.21 when GL was modeled as a cow phenotype, and from 0.14 to 0.69 when modeled as a fetus phenotype. The model including fetal additive genetic, maternal, and paternal effects, along with maternal and paternal permanent environmental effects, showed the best performance based on DIC and predictive ability, achieving the highest accuracy (0.38) and lowest bias (0.55). Rank correlations between this model and others ranged from 0.46 to 0.99. Incorporating genomic information increased accuracy and dispersion, reduced bias, and yielded higher correlations between genomic estimated breeding values (GEBVs) using complete and reduced datasets. Furthermore, genome-wide association studies (GWAS) for GL identified genomic regions associated with reproductive processes. Overall, modeling GL as a fetus phenotype provided a better fit and predictive capacity for genetic evaluation in Holstein cattle. The inclusion of genomic information, together with the identification of genomic regions associated with GL, further reinforces its importance in improving genetic evaluations.

Keywords: 2026

How to Cite:

Abreu, C., Costa, C., Alves Silva, D., Veroneze, R., Marques, D., Carvalheira, J., Lopes, P. & Silva, A. A., (2026) “Genetic evaluation of gestation length in Holstein cattle: comparison between fetal and dam genetic effects.”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286983. doi: https://doi.org/10.31274/wcgalp.24187

Rights: 1