Efficacy of plant breeding using genomic information

Area of Science:

• Agricultural Science
• Genetics
• Quantitative Genetics

Background:

• Genomic selection (GS) has revolutionized plant and animal breeding over the past two decades.
• Despite widespread adoption, various factors influence the efficacy of GS.
• Understanding the impact of genomic information on prediction accuracy is crucial for optimizing breeding strategies.

Purpose of the Study:

• To quantitatively assess the increase in prediction accuracy when utilizing genomic information compared to traditional methods.
• To investigate the influence of marker quality and individual relatedness on prediction gains.
• To validate the importance of genomic data for enhancing realized genetic gain in breeding programs.

Main Methods:

• Analysis of 14 real-world datasets encompassing diverse traits and environments.
• Comparison of prediction accuracy metrics with and without the inclusion of genomic data.
• Evaluation of the correlation between marker density, individual relatedness, and prediction accuracy gains.

Main Results:

• Genomic information inclusion resulted in an average prediction accuracy gain of 26.31% across datasets, traits, and environments.
• Specific metrics showed substantial improvements: 46.1% gain in Pearson's correlation and 6.6% gain in normalized root mean squared error.
• Higher marker quality and individual relatedness correlated with greater prediction accuracy gains.

Conclusions:

• Genomic information is vital for improving prediction accuracy in breeding programs.
• The magnitude of accuracy gain is influenced by marker quality and the relatedness of individuals.
• GS significantly contributes to realizing greater genetic gain in genomic-assisted breeding.