Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: BIOINFORMATIC METHODS AND TOOLS TO PREDICT SMALL GRAIN FIELD PERFORMANCE Title: Dynamics of long-term genomic selection

Author
item Jannink, Jean-Luc

Submitted to: Genetics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: April 25, 2010
Publication Date: August 16, 2010
Citation: Jannink, J. 2010. Dynamics of long-term genomic selection. Genetics. 42:35.

Interpretive Summary: We know from simulation and empirical studies that a method called genomic selection (GS) can predict, using DNA marker data, the performance of breeding lines for quantitative traits with sufficient accuracy that selection on predictions could generate rapid gains in early selection cycles. Beyond those cycles, many mechanisms make it difficult to predict what longer-term selection gains will be. The best way to make such predictions is through simulating the breeding process. Starting from marker data on 192 breeding lines from an elite six-row spring barley program, we simulated several scenarios under which GS might be used to determine how well it might do in the long term, to understand what mechanisms might most affect its performance, and to propose approaches to improve that performance. Assuming that applying GS prior to phenotyping shortened breeding cycle time by 50%, that practice strongly increased early selection gains but also caused the loss of many favorable alleles, leading to low long-term gain. Placing additional weight on low-frequency favorable marker alleles, however, led to higher long-term gains without large losses in short-term gains.

Technical Abstract: Simulation and empirical studies of genomic selection (GS) show accuracies sufficient to generate rapid gains in early selection cycles. Beyond those cycles allele frequency changes, recombination, and inbreeding make analytical prediction of gain impossible. On the basis of marker data on 192 breeding lines from an elite six-row spring barley program, stochastic simulation was used to explore the effects of large or small initial training populations, heritabilities of 0.2 or 0.5, applying GS before or after phenotyping, and applying additional weight on low-frequency favorable marker alleles. Genomic predictions were from ridge regression or a Bayesian analysis. Assuming that applying GS prior to phenotyping shortened breeding cycle time by 50%, that practice strongly increased early selection gains but also caused the loss of many favorable QTL alleles, leading to loss of genetic variance, loss of GS accuracy, and a low selection plateau. Placing additional weight on low-frequency favorable marker alleles, however, allowed GS to increase their frequency early on, causing an initial increase in genetic variance. This dynamic led to higher long-term gains while mitigating losses in short-term gains. Weighted GS also increased the maintenance of marker polymorphism, ensuring that QTL-marker linkage disequilibrium was higher than in unweighted GS.

Last Modified: 11/26/2014