Access Official Technical PDF

Share Summary:LinkedInEmail

Genomic selection is a plant-breeding approach that uses genome-wide marker data to estimate which plants are most likely to pass desirable traits to the next generation. Unlike approaches that focus only on markers already associated with a specific trait, genomic selection uses information from markers across the genome, whether or not each marker’s association with the trait has been established as significant.

For breeding teams managing many lines and traits, the approach offers a way to use genetic data to guide selection. Its practical value depends on how well the model reflects the breeding population and how effectively marker and phenotype information are assembled and applied.

How genomic selection works in crop improvement

A genomic selection scheme relies on two related populations: a training population and a breeding population. The training population provides the information used to build the selection model. Its marker data and observed trait information are used to estimate marker-associated effects. Those estimates can then be used to calculate genomic estimated breeding values, or GEBVs, for individual plants or lines in the breeding population.

The breeding population is the group being evaluated for improvement. Breeders use its GEBVs to help select superior lines for further breeding or for use as parents in the development of improved hybrids. In simplified terms, the training population establishes how genomic information relates to breeding value, while the breeding population is where that relationship is used to guide selection.

This distinction is important: genomic selection does not mean that phenotypic information is irrelevant. The training process depends on trait observations, while genome-wide marker data extend the basis for estimating breeding value to the individuals being considered for selection.

Genomic selection and marker-assisted selection

Marker-assisted selection typically relies on markers linked to known genes or quantitative trait loci associated with a trait. Genomic selection takes a broader approach. It uses marker information distributed across the genome without requiring each marker to have a previously confirmed, significant association with the target trait.

That difference can reduce reliance on identifying and mapping specific QTLs before using marker information in a selection scheme. Rather than asking only whether a particular marker is linked to a known genetic factor, genomic selection combines information across many markers to estimate a plant’s overall breeding value.

Genomic selection is therefore best understood as a development in marker-based breeding, not a replacement for every other genetic or phenotypic method. Its results depend on the quality and relevance of the data used to train the model and on how well the approach fits the breeding objectives.

Potential benefits for breeding programs

One potential advantage is the opportunity to improve selection efficiency, including for traits with low heritability, where observed phenotype alone may provide a less reliable basis for choosing plants. The approach can also reduce the need to phenotype every selection cycle in the breeding population. This may be especially useful where breeding cycles are long, as in perennial species.

By using GEBVs, breeding teams may also consider lines that have not been evaluated in the target environment when selecting parents for hybridization. The method can incorporate marker and trait information gathered over time from different evaluation programs and environments, supporting a broader evidence base for selection.

The approach may also help manage the trade-off between selection gains and the maintenance of genetic variability. The underlying breeding discussion suggests that genomic selection could reduce inbreeding and loss of variability compared with selection based on breeding values estimated from phenotype data, without sacrificing selection gains. These outcomes are not automatic; they depend on how selection is designed and applied.

Implementation considerations and limitations

Genomic selection’s benefits should be weighed against its limitations. One concern is that the algorithms may not explain the biological processes underlying a phenotype. Because selection is based on estimated effects from genome-wide data, the approach may not make direct use of all findings from genetics and genomics, such as identified QTLs and the genes they represent.

This creates an opportunity to develop models that combine genomic selection with biological knowledge. The aim would be to retain the broad use of marker data while making better use of what is known about the genetics of the traits being improved.

Adoption also depends on resolving practical issues within each breeding program. The training population must provide useful information for the breeding population, and the resulting estimates need to be appropriate for the selection decisions being made. The approach’s value should be assessed in relation to the program’s traits, populations, environments, and breeding goals.

What breeding teams should take away

Genomic selection provides a framework for turning genome-wide marker data into estimated breeding values that can guide crop improvement. Its central workflow—train a model with one population, then apply marker-effect estimates to another—offers a route to selection without requiring prior discovery of significant associations for every marker used.

For breeders and technical decision-makers, the practical implication is to treat genomic selection as a data-driven component of a wider breeding strategy. It can support selection, potentially reduce phenotyping requirements in some cycles, and help teams use accumulated information across evaluations. Its effectiveness still depends on sound training data, appropriate application, and continued attention to biological understanding. As genotyping costs decline and implementation challenges are addressed, genomic selection is positioned to become a more widely integrated tool in plant-breeding programs.

Access Official Technical PDF

About The Author