# Unraveling the genetic basis of general combining ability in CIMMYT elite bread wheat germplasm: implications for breeding strategies optimization

## TL;DR

This study identifies genetic markers and biological pathways linked to wheat breeding efficiency, aiming to improve cross selection strategies.

## Contribution

The study provides new insights into genetic markers and pathways influencing wheat combining ability through GWAS, GO analysis, and genomic prediction.

## Key findings

- Thirteen marker-trait associations were identified for grain yield, a combined index, and progeny number per-cross.
- Genes related to hydrogen peroxide metabolism and oxidative stress response were highlighted as important for the studied traits.
- The RKHS model achieved the highest prediction accuracy for grain yield at 0.34.

## Abstract

In wheat breeding programs, several hundred crosses are performed annually, but only individuals from a few families advance to the final stages of the breeding pipelines. Therefore, a deeper understanding of the general combining ability (GCA) of wheat genotypes might enhance the breeding efficiency in selecting parents. For this reason, we tested the performance of the offspring of ~1200 parental elite lines. Using a genome-wide association study (GWAS), gene ontology (GO) analysis, and genomic prediction (GP), our objectives were to i) identify marker-trait associates (MTAs) and candidate genes, ii) assess temporal allele frequency dynamics of identified MTAs, and iii) estimate prediction accuracy (PA) for key traits: Progeny Number per-Cross (PNC), grain yield (GY), and a combined index incorporating these traits (“index”). Our findings revealed a total of 13 MTAs: eight for GY, four for the “index”, and one for PNC. The GO analysis highlighted several genes involved in hydrogen peroxide metabolism and catabolism processes (H2O2), reactive oxygen species, response to oxidative stress, cell wall biogenesis, the metabolic process of modified amino acids at the cellular level, and glutathione metabolic process for the studied traits. Notably, allele frequency analysis over time indicated that most MTAs are under positive selection, likely reflecting indirect breeder-driven selection. The highest PA was reached by using the reproducing kernel Hilbert space (RKHS) model for the trait GY (0.34). The identification of MTAs for PNC and GY provided insight into the biological pathways underpinning combining ability and demonstrated the potential for predicting the ability of the genotypes to be crossed. These findings might contribute to the optimization crossing strategy saving costs and increasing the breeding program efficiency.

Illustration divided into three panels. The left panel labeled “GWAS” shows a plant with various colored circles labeled Gy, Index, and PNC, with roots connected to colors and numbers. The middle “Gene ontology” panel features a plant with connections to biological processes such as “Reactive oxygen species metabolic process” and “Response to stress”. The right “Genomic Prediction” panel displays a plant with colored circles labeled Gy, Index, and PNC, alongside numerical values next to GBLUP and RKHS, indicating prediction values.