Initial findings on genomic selection indicated substantial improvement in major traits, such as performance. Under genomic selection, the selection accuracy increases, and the generation interval decreases, accelerating the selection response. However, recent unofficial reports indicate an increased frequency of deterioration of unselected or negatively correlated traits. This phenomenon may arise due to the mismatch between the accelerated gains and the overlooked changes in correlated traits. Because of the rapid turnover of genomic selection, heritabilities for production traits decline faster, and the genetic antagonism between production and fitness traits intensifies. Therefore, it is crucial to look for unexpected changes in economically important traits and take rapid steps to prevent further declines, especially in secondary traits. However, estimating variance components and genetic parameters over time to investigate such changes with many genotyped animals is prohibitive. Without that, assessing and preventing the negative impact of genomic selection becomes challenging. Therefore, we propose to:
1) Extend the limits of current methods to estimate variance components with large genomic datasets;
2) Develop new methods to estimate genetic correlations and heritabilities based on crossvalidation equations that use predictive ability or linear regression methods, which will work with millions of genotyped animals;
3) Test the extended/new methods on various datasets to identify the combination of traits with anticipated negative effects of genomic selection.
This project will provide tools for the US industry to identify and avoid unexpected changes due to genomic selection, which aligns well with the priorities of Program A1201 – “Animal Breeding, Genetics, and Genomics.”
Funder: USDA NIFA
Amount: $650,000
PI: Daniela Lino Lourenco, College of Agricultural and Environmental Sciences, Department of Animal and Dairy Sciences