Scholarlink

<p>Analytical methods that explore genotype-by-environment interaction (GEI) in multi-environment trials (MET) are essential for the conduction of segregating populations and the release of tropical wheat (<i>Triticum aestivum L</i>.) cultivars. Linear mixed models not only deal with data imbalance, but also allow for the modeling of (co)variances (VCOV), making use of the biological process of GEI for the selection of superior genotypes. The objective of this study was to compare different modeling structures for residual and genetic VCOV in the selection of the best tropical wheat populations for generation advancing in a breeding program. Initially, the dataset was composed of 25 segregating populations evaluated for grain yield (GY) in MET, involving six tropical environments in the F<sub>2</sub>, F<sub>3</sub> and F<sub>4</sub> generations. However, although it started with 25 populations, the actual number evaluated varied depending on the environment and generation. The model that considered heteroscedasticity and different genetic covariances in the genotypic effects, in addition to heteroscedasticity with null covariances for the residual effects, had the lowest values for the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) compared to the baseline model. Based on these results, the populations P7/P2, P9/P1, P7/P1, P9/P2, P7/P4 and P8/P1 showed greater potential to derive superior lines of tropical wheat. In addition to enabling the selection of the best populations, the chosen model allows the effective selection of superior populations in the initial generations.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Statistical genetics models with residual and genetic structures enhance the accuracy of selecting wheat populations

  • 期刊论文

  • 发表日期：

    2025年4月9日

查看全文
  • João Marcos Amario de Sousa,
  • Eduardo Filipe Torres Vieira,
  • Rafael Tobias Lang Fronza,
  • João Amarro Ferreira Vieira Netto,
  • Gabriel Wolter Lima,
  • Kaio Olimpio das Graças Dias,
  • Maicon Nardino

摘要

Analytical methods that explore genotype-by-environment interaction (GEI) in multi-environment trials (MET) are essential for the conduction of segregating populations and the release of tropical wheat (Triticum aestivum L.) cultivars. Linear mixed models not only deal with data imbalance, but also allow for the modeling of (co)variances (VCOV), making use of the biological process of GEI for the selection of superior genotypes. The objective of this study was to compare different modeling structures for residual and genetic VCOV in the selection of the best tropical wheat populations for generation advancing in a breeding program. Initially, the dataset was composed of 25 segregating populations evaluated for grain yield (GY) in MET, involving six tropical environments in the F2, F3 and F4 generations. However, although it started with 25 populations, the actual number evaluated varied depending on the environment and generation. The model that considered heteroscedasticity and different genetic covariances in the genotypic effects, in addition to heteroscedasticity with null covariances for the residual effects, had the lowest values for the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC) compared to the baseline model. Based on these results, the populations P7/P2, P9/P1, P7/P1, P9/P2, P7/P4 and P8/P1 showed greater potential to derive superior lines of tropical wheat. In addition to enabling the selection of the best populations, the chosen model allows the effective selection of superior populations in the initial generations.