Genomic prediction with haplotype blocks in wheat.

Area of Science:

• Plant breeding
• Quantitative genetics
• Bioinformatics

Background:

• Genomic prediction utilizes marker data to predict breeding values.
• Haplotype blocks, encompassing multiple single nucleotide polymorphisms (SNPs), may offer advantages over individual SNPs for genomic prediction.
• Optimal methods for constructing haplotype blocks for enhanced prediction accuracy remain underexplored.

Purpose of the Study:

• To compare the prediction accuracies of various haplotype block construction methods against single SNP prediction in winter wheat.
• To evaluate the performance of different genomic prediction models (RR-BLUP, RMLA, GBLUP) using haplotype blocks.
• To identify the most effective haplotype block strategies for predicting 11 different traits in winter wheat.

Main Methods:

• Haplotype blocks were constructed using linkage disequilibrium, fixed SNP numbers, fixed lengths, and the HaploBlocker R package.
• Genomic prediction was performed using RR-BLUP, RMLA, and GBLUP models with cross-validation on data from 361 winter wheat lines.
• Prediction accuracies were assessed for 11 different traits, including disease resistance and plant height.

Main Results:

• Linkage disequilibrium-based haplotype blocks yielded the highest accuracies for resistance to B. graminis, P. triticina, and F. graminearum.
• Blocks defined by fixed marker numbers or lengths in centimorgans were superior for predicting plant height.
• HaploBlocker-derived blocks outperformed other methods for protein concentration and resistance to S. tritici, B. graminis, and P. striiformis.

Conclusions:

• The optimal method for constructing haplotype blocks for genomic prediction is trait-dependent.
• Haplotype blocks can capture complex genetic effects like epistasis and improve prediction accuracy for specific traits.
• Further research is needed to optimize haplotype block characteristics and modeling approaches to maximize their utility in genomic selection.