Genomics Assisted Breeding Strategy in Flax

Genomic selection (GS) or genomic prediction (GP) is a type of marker-assisted selection that relies on genome-wide markers to predict genomic-estimated breeding values (GEBVs) of phenotypes. GS is quickly becoming a conventional approach in both plant and animal breeding to increase selection accuracy, reduce breeding cost and shorten breeding cycles. The concept of GS models was first developed using genome-wide random markers, with marker density being a key element in estimating the predictive ability in breeding populations. It is currently straightforward to generate high-density marker datasets thanks to the remarkable advances in genotyping technologies. Recent studies showed that high-density genome-wide random markers do not necessarily generate high genomic predictive ability in GS because the vast majority of markers are unrelated to the traits of interest, thus generating background noises and lowering the predictive ability. Alternatively, the use of quantitative trait loci (QTLs), identified through genome-wide association study (GWAS) methods, in GS models can significantly improve genomic predictive ability and reduce the genotyping cost of the test populations. Here, we present recent findings, discuss a few case studies, a QTL-based GS strategy and a genomic cross-predictions for flax breeding improvement.