Factors affecting accuracy from genomic selection in populations derived from multiple inbred lines: a Barley case...

Area of Science:

• Quantitative genetics
• Plant breeding
• Genomic selection

Background:

• Genomic selection (GS) uses marker data to predict breeding values, aiding crop and livestock improvement.
• Accurate prediction requires understanding how factors like marker density, linkage disequilibrium (LD), and population structure influence GS methods.

Purpose of the Study:

• To compare the accuracy of four genomic selection prediction methods.
• To evaluate the impact of marker density, LD levels, quantitative trait locus (QTL) number, sample size, and replication on prediction accuracy.

Main Methods:

• Simulated high and low LD populations from 42 barley inbred lines.
• Simulated true breeding values (TBV) based on 20 or 80 additive QTL.
• Compared four methods: RR-BLUP, Bayes-B, Bayesian shrinkage regression, and marker-based relationship BLUP (MACE).

Main Results:

• Genomic estimated breeding value (GEBV) accuracies were comparable to phenotypic accuracies, reducing field evaluation needs.
• A trade-off exists between capturing marker-QTL LD and marker-based relatedness.
• Marker-based relationship BLUP (MACE) generally outperformed marker effect methods, highlighting the value of relationship information.
• Bayes-B captured both LD and relatedness, while Bayesian shrinkage focused on LD and MACE on relationships.

Conclusions:

• Marker-based relationship BLUP (MACE) is highly effective for genomic prediction, especially in diverse populations.
• The choice of method depends on population structure, LD, and QTL characteristics.
• Bayes-B offers a balanced approach, capturing both LD and relatedness for robust genomic prediction.