Genome-wide association analysis accounting for genotype by diet interaction for feed efficiency in Holstein cows

Authors: AMBIKE, V - Michigan State University; VANDEHAAR, M - Michigan State University; PARKER-GADDIS, K - Council On Dairy Cattle Breeding; PEANGARICANO, F - University Of Wisconsin; WHITE, H - University Of Wisconsin; WEIGEL, K - University Of Wisconsin; Baldwin, Ransom; SANTOS, J - University Of Florida; KOLTES, JAMES - Iowa State University; TEMPELMAN, R - Michigan State University

Submitted to: World Congress of Genetics Applied in Livestock Production
Publication Type: Proceedings
Publication Acceptance Date: 4/20/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: Feed efficiency (FE) characterizes a dairy cow’s ability to convert feed into milk relative to her intake and maintenance requirement, thereby impacting economic and environmental sustainability. Key FE traits include dry matter intake (DMI) and residual feed intake (RFI), the latter being the difference between actual and predicted DMI (Connor, 2015). Dietary components like fat, starch, neutral detergent fiber (NDF), and crude protein (CP) may influence the genetic expression of DMI and RFI, thereby making the assessment of genotype-by-diet (G×D) interactions potentially important for precision genetic management of livestock. Random regression models (RRM) model genetic merit as linear or higher order functions of continuous environmental gradients (e.g., temperature). Genome-wide association (GWA) studies are used to identify potentially important genomic regions associated with complex traits by inferring statistical associations between phenotypes and individual single nucleotide polymorphisms (SNP) across the genome. In BLUP-GWA, estimated SNP effects are backsolved from genomic prediction models, which consider all markers simultaneously, thereby improving detection of quantitative trait loci (QTL) (Aguilar et al., 2019). Some G×D interactions can be due to scaling, whereby animals rank similarly across different environments albeit with heterogeneous variability, whereas some interaction is due to reranking of animals across environments thereby impacting selection programs. Extending the GWA framework to account for G×D potentially allows the detection of genomic regions that are robust versus sensitive to different dietary components for feed efficiency. These inferences could facilitate selection of robust animals, and discovery of putative candidate genes for robustness. Thus, the objective of this study was to conduct a GWA study for DMI and RFI within a RRM framework for modeling GxD.