Estimation & Prediction

Indirect Genomic Predictions for Calving Traits in Crossbred Holstein-Jersey dairy cattle

---

Abstract

Single-step GBLUP (ssGBLUP) enables the derivation of SNP effects for indirect prediction (IP) of genomic breeding values (GEBV) in young genotypes animals by backsolving GEBVs. This method offers a practical and efficient solution for genomic evaluations involving large numbers of genotyped animals, facilitating more frequent evaluations at reduced cost. The study aimed to predict genomic breeding values for calving traits in crossbred dairy cattle using ssGBLUP. The traits analyzed included calving ease (CE), gestation length (GL), and stillbirth (SB), with phenotypic records ranging from 4,408,115 (SB) to 8,812,994 (GL). Pedigree data from 8,443,224 animals supported breeding value estimation through multiple-trait model, capturing the genetic correlations among these traits to enhance prediction accuracy. Genotypes from 955,053 Holstein, 268,082 Jersey, and 28,979 Holstein-Jersey crossbred animals were imputed to 45K SNP markers selected based on their call rates and minor allele frequencies to ensure data quality. The Algorithm for Proven and Young (APY) was applied using 30,000 core animals randomly selected to generate the benchmark genomic estimated breeding values (GEBV). This method efficiently handles large genomic datasets by approximating the inverse of the genomic relationship matrix, enabling scalable analyses. SNP effects were estimated from a multibreed evaluation incorporating all Holstein, Jersey, and crossbred animals. Indirect predictions of the validation animals (576 crossbred animals) were calculated using the resulting SNP effects. The predictive ability of IP was calculated as the Pearson correlation between IP and GEBV of the validation animals. Correlations between predicted transmitting ability (PTA) obtained using regular ssGBLUP and IP approached 0.84 for CE, 0.88 for SB, and 0.94 for GL. These high correlations indicate that IP provides reliable genomic predictions for crossbred animals. Overall, these results demonstrate that indirect predictions for crossbred animals can be derived using ssGBLUP when imputed purebred genotypes are included in the evaluation. This approach offers a cost-effective and scalable strategy to enhance genomic selection programs, particularly in populations with complex breed compositions. Future work may focus on integrating additional functional genomic data and exploring the impact of different imputation strategies to further improve prediction accuracy.

Keywords: 2026

How to Cite:

Vargas, G., Pacheco, H., Vukasinovic, N., Sánchez-Castro, M., Oliveira, G., Passafaro, T. & Gonzalez-Peña, D., (2026) “Indirect Genomic Predictions for Calving Traits in Crossbred Holstein-Jersey dairy cattle”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285722. doi: https://doi.org/10.31274/wcgalp.23782

Rights: 1