Volume 20, Issue 2 (1-2026)
MGj 2026, 20(2): 0-0 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Jahromi M, Abbasi A, Bihamta M R, Armion M. Prediction accuracy improvement of genomic selection models through the optimization of models using covariates in bread wheat (Triticum aestivum L.). MGj 2026; 20 (2) : 3
URL: http://mg.genetics.ir/article-1-1874-en.html
URL: http://mg.genetics.ir/article-1-1874-en.html
Abstract: (37 Views)
Genomic prediction has emerged as a promising approach to accelerate the genetic improvement of complex traits in wheat breeding. However, improving the prediction accuracy of genomic prediction models remains a major challenge for their effective implementation. In this context, the appropriate use of covariates has shown considerable potential for improving model performance. In this study, two covariates—days to heading (DH) and days to flowering (DF)—were used to evaluate their contribution to improving the prediction accuracy of seven genomic selection (GS) models, including GBLUP, RRBLUP, and Bayesian models (BayesA, BayesB, BayesC, BayesRR, and BayesLASSO). Accordingly, two strategies were employed: GS models without covariates (strategy one) and GS models incorporating covariates (strategy two). The experiment was conducted at the University of Tehran research field using 100 genotypes across two cropping seasons (2019–2020 and 2020–2021). The covariates exhibited strong phenotypic and genotypic correlations with each other and showed significant positive and negative associations with the studied traits. Trait heritability ranged from 69% for green fluorescence protein (GFP) to 97% for DH. Genomic predictions were performed using 65,948 SNP markers in combination with phenotypic data. By incorporating DH and DF as covariates, the highest prediction accuracy for grain yield (GY) increased from 0.085 under strategy one to 0.279 under strategy two. Similar improvements were observed for thousand-kernel weight (TKW) (0.238 to 0.411), biological yield (BY) (0.230 to 0.399), GFP (0.314 to 0.714), and days to physiological maturity (DFM) (0.256 to 0.889). Under strategy two, Bayesian models—particularly BayesRR and BayesLASSO—exhibited the highest prediction accuracies. These findings indicate that incorporating appropriate covariates into GS models is an effective strategy for optimizing model performance and enhancing the genomic prediction accuracy of key phenological and agronomic traits in wheat.
Article number: 3
Type of Study: Applicable |
Subject:
Subject 01
Received: 2024/12/9 | Accepted: 2026/02/1 | Published: 2026/02/3
Received: 2024/12/9 | Accepted: 2026/02/1 | Published: 2026/02/3
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
