Scholarlink

<p>Cultivar recommendation is a critical stage in plant breeding programs, and selecting superior varieties for multiple traits remains a challenge due to the variety <i>×</i> environment (V<i>×</i>E) interaction. The Pan-African Trials Network aims to expand soybean cultivar recommendations across new tropical and subtropical regions. However, varieties exhibit distinct adaptation patterns across environments. Bayesian probabilistic models offer a way to manage cultivar recommendation risk through the V<i>×</i>E interaction. This study aimed to identify soybean varieties with high probabilities of superior performance across years and locations using a Bayesian Probabilistic Selection Index (BPSI) in a multi-trait and multi-environment framework. Ninety-seven soybean varieties were evaluated across 37 environments and six crop seasons in Zimbabwe using a randomized complete block design with three replications. Traits evaluated included grain yield, plant height, and lodging. The probability of superior performance was estimated using a 20% selection intensity for each trait and 10% for the multi-trait index. Plant height showed the highest experimental precision, whereas grain yield exhibited greater variability across environments. Most varieties performed better across groups of years than groups of locations, indicating stronger variety <i>×</i> location than variety <i>×</i> year interaction. The BPSI identified ten varieties with high probabilities of superior performance across environments in Zimbabwe. Selected varieties achieved probabilities above 60% for grain yield, 80% for plant height, and 50% for lodging, indicating greater yield stability and reduced selection risk. Pairwise comparisons confirmed that unselected varieties performed worse than selected ones. By integrating the probability of superior performance within and across environments, the BPSI effectively identified varieties with both specific and broad adaptation. This Bayesian framework provides a robust, data-driven approach for cultivar recommendation in Zimbabwe, combining predictive accuracy and risk management. The BPSI shows strong potential to enhance soybean breeding decisions under heterogeneous environmental conditions, and further studies should extend this approach to other crops, traits, with enviromic and genomic data.</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

Leveraging probabilistic models to enhance soybean cultivar recommendation in Zimbabwe

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月4日

查看全文
  • José Tiago B. Chagas,
  • Mauricio S. Araújo,
  • Marcela Martinez,
  • João P.S. Pavan,
  • Learnmore Mwadzingeni,
  • Ronica Mukaro,
  • Andrew Henderson,
  • Charles Mutimaamba,
  • Andrew Scaboo,
  • Erica P. Leles,
  • Michelle F. Santos,
  • Brian W. Diers,
  • Peter Goldsmith,
  • José Baldin Pinheiro

摘要

Cultivar recommendation is a critical stage in plant breeding programs, and selecting superior varieties for multiple traits remains a challenge due to the variety × environment (V×E) interaction. The Pan-African Trials Network aims to expand soybean cultivar recommendations across new tropical and subtropical regions. However, varieties exhibit distinct adaptation patterns across environments. Bayesian probabilistic models offer a way to manage cultivar recommendation risk through the V×E interaction. This study aimed to identify soybean varieties with high probabilities of superior performance across years and locations using a Bayesian Probabilistic Selection Index (BPSI) in a multi-trait and multi-environment framework. Ninety-seven soybean varieties were evaluated across 37 environments and six crop seasons in Zimbabwe using a randomized complete block design with three replications. Traits evaluated included grain yield, plant height, and lodging. The probability of superior performance was estimated using a 20% selection intensity for each trait and 10% for the multi-trait index. Plant height showed the highest experimental precision, whereas grain yield exhibited greater variability across environments. Most varieties performed better across groups of years than groups of locations, indicating stronger variety × location than variety × year interaction. The BPSI identified ten varieties with high probabilities of superior performance across environments in Zimbabwe. Selected varieties achieved probabilities above 60% for grain yield, 80% for plant height, and 50% for lodging, indicating greater yield stability and reduced selection risk. Pairwise comparisons confirmed that unselected varieties performed worse than selected ones. By integrating the probability of superior performance within and across environments, the BPSI effectively identified varieties with both specific and broad adaptation. This Bayesian framework provides a robust, data-driven approach for cultivar recommendation in Zimbabwe, combining predictive accuracy and risk management. The BPSI shows strong potential to enhance soybean breeding decisions under heterogeneous environmental conditions, and further studies should extend this approach to other crops, traits, with enviromic and genomic data.