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 #402339
Research Project: Improving Dairy Cow Feed Efficiency and Environmental Sustainability Using Genomics and Novel Technologies to Identify Physiological Contributions and Adaptations

Location: Animal Genomics and Improvement Laboratory

Title: Pedigree and genomic adjustments for single-step genomic BLUP applied to residual feed intake

Author
item LEGARRA, ANDRES - Council On Dairy Cattle Breeding
item VANDEHAAR, MICHAEL - Michigan State University
item TEMPELMAN, ROBERT - Michigan State University
item KOLTES, JAMES - Iowa State University
item WHITE, HEATHER - University Of Wisconsin
item WEIGEL, KENT - University Of Wisconsin
item Baldwin, Ransom - Randy
item Vanraden, Paul
item PENAGARICANO, FRANCISCO - University Of Wisconsin
item SANTOS, JOSE - University Of Florida
item PARKER GADDIS, KRISTEN - Council On Dairy Cattle Breeding

Submitted to: Journal of Dairy Science
Publication Type: Abstract Only
Publication Acceptance Date: 3/21/2023
Publication Date: 6/25/2023
Citation: Legarra, A., Vandehaar, M., Tempelman, R., Koltes, J., White, H., Weigel, K., Baldwin, R.L., Van Raden, P.M., Penagaricano, F., Santos, J., Parker Gaddis, K. 2023. Pedigree and genomic adjustments for single-step genomic BLUP applied to residual feed intake [abstract]. Journal of Dairy Science. 106(Suppl. 1):189(abstr. 2714).

Interpretive Summary:

Technical Abstract: Use of VanRaden’s G requires, in principle, use of allele frequencies at the foundation of pedigree (“base allele frequencies”). Use of ssGBLUP implies that genomic relationships of non-genotyped individuals are predicted from G using pedigree relationships in A. This requires that both G and A describe relationships from the same base population, a requisite often called “compatibility”. Three major methods to achieve compatibility are: (1) use of base allele frequencies (“base”) (2) use of observed allele frequencies across all genotyped animals, followed by a correction to force same average BV and same genetic variance (“tuning”) and (3) use of (fictitious) 0.5 allele frequencies followed by corrections of pedigree relationships to match observed genomic relationships through the theory of metafounders (“MF”). In theory the best alternatives are “base” for single base populations if these are known, and “MF” for multiple base populations or if base allele frequencies are difficult to estimate. Here we compare the three options in a subset of 9K Holstein individuals phenotyped for Residual Feed Intake, and a corresponding pedigree of 56K individuals, of whom 13K with genotypes. Base allele frequencies were estimated from old genotyped animals using findhap.f90. We modelled a single base population or metafounder, with gamma=0.235 estimated by Maximum Likelihood. The correlations (r) and regressions (b) of genomic relationships G on pedigree relationships A are shown in Table 1. The best option seems “MF” followed by “base” and then by “tuning”. No method has regression slopes of 1 as might be expected. Correlations among GEBVs were: “base”-“tuning”: 0.89, “base”-“MF”: 0.998, “observed”-“MF”: 0.90, showing that the resulting GEBVs do differ depending on the different choices although, in this data set, “base” and “MF” resulted in nearly identical results. Heritability estimates were 0.16 and nearly identical in all cases.