Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Animal Genomics and Improvement Laboratory » Research » Publications at this Location » Publication #412959
Research Project: Increasing Accuracy of Genomic Prediction, Developing Algorithms, Selecting Markers, and Evaluating New Traits to Improve Dairy Cattle

Location: Animal Genomics and Improvement Laboratory

Title: Multibreed single-step genomic best linear unbiased predictor evaluations for fertility traits in US dairy cattle

Author
item TABET, JOE-MENWER - University Of Georgia
item LOURENCO, DANIELA - University Of Georgia
item BERMANN, MATIAS - University Of Georgia
item MISZTAL, IGNACY - University Of Georgia
item Vanraden, Paul
item LEGARRA, ANDRES - Council On Dairy Cattle Breeding

Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 5/21/2024
Publication Date: 6/25/2024
Citation: Tabet, J., Lourenco, D., Bermann, M., Misztal, I., Van Raden, P.M., Legarra, A. 2024. Multibreed single-step genomic best linear unbiased predictor evaluations for fertility traits in US dairy cattle [abstract]. Journal of Dairy Science. 107(Suppl. 1):130(abstr. 1516).

Interpretive Summary:

Technical Abstract: We tested ssGBLUP for fertility traits in realistic conditions, mimicking the CDCB December 2022 official runs. We included all historical phenotypes for Daughter Pregnancy Rate (DPR), Cow Conception Rate (CCR), Heifer Conception Rate (HCR), and Early First Pregnancy (EFC), comprising 94 million records. The full pedigree consisted of 94 million individuals across the six official breeds and their crosses, and 417 Unknown Parent Groups (UPG) modeled as unrelated or metafounders. We used about 2 million genotyped animals with records or descendants with records in ssGBLUP with the algorithm for proven and young (APY). A total of 45K genotyped animals representing the six breeds and crosses were set as core. Results from the ssGBLUP run were compared with those from BLUP regarding dispersion bias and prediction reliability. For that, the whole dataset contained phenotypes up to December 2022, and the partial dataset was truncated in December 2018. Then, we compared (G)EBV of bulls with no daughters with records in the partial and at least 100 daughters with DPR records in the whole dataset, resulting in 1891 bulls for Holstein and 303 for Jersey. The linear regression method (LR) was used for validation. Genetic trends were also investigated, which showed a steep decrease until about 2000–2010, followed by a fast increase after including fertility in the genetic evaluation and selection index and using genomic information. The regression coefficients (b1) varied between 0.89 and 0.94 for ssGBLUP with metafounders and between 0.88 and 0.90 with UPG, resulting in a slightly biased evaluation. For BLUP, the values were from 0.76 to 0.93 and 0.72. to 0.92, respectively, which is more biased because BLUP cannot account for the genomic selection of young animals. The correlation between early (partial dataset) and late (whole dataset) proofs from BLUP varied from 0.49 to 0.65 for metafounders and 0.50 to 0.63 for UPG, whereas for genomic proofs, they varied between 0.80 and 0.91 with metafounders and 0.81 to 0.89 with UPG, showing good reliability of early genomic proofs. Overall, ssGBLUP seems viable for the genomic evaluation of fertility traits in US dairy cattle.