Scholarlink

<p>Because climate change is pressuring every terrestrial ecosystem including forests, efficient methods for managing forest genetic resources are needed. One approach is to use molecular genetic markers in tree breeding programmes. Current genotyping tools can be used to improve the selection of genetically superior trees via genomic prediction. However, population structure may affect the prediction of breeding values, particularly in the early stages of tree domestication. We examined how models of population structure affected breeding values using a single-step genomic prediction framework (ssBLUP) in Douglas-fir (<i>Pseudotsuga menziesii</i> (Mirbel) Franco) in New Zealand. In particular, we tested three metafounder scenarios. The ssBLUP1 scenario assumed there was a single metapopulation. The other scenarios assumed there were multiple metapopulations with either no gene flow (ssBLUP2) or gene flow among populations (ssBLUP3). All three metafounder scenarios performed better than the pedigree-based (ABLUP) and benchmark (ssBLUP) scenarios. The multiple metapopulation scenarios (ssBLUP2 and ssBLUP3) had slightly greater heritabilities than did the ABLUP and ssBLUP approaches, but heritabilities doubled for the single metapopulation scenario (ssBLUP1). The ssBLUP1 scenario also had the largest accuracy of genomic estimated breeding values (GEBV) but also the largest GEBV bias. In contrast, ssBLUP3 had the lowest GEBV accuracy and bias. Results for the ssBLUP2 scenario demonstrated the value of accounting for genetic variation among metafounders. Compared to the single metapopulation scenario (ssBLUP1), ssBLUP2 had lower GEBV bias and smaller underdispersion of GEBV. Finally, our study demonstrated the benefits of selecting ancestry-informative markers (AIM) for use in the ssBLUP model with metafounders. Compared to all markers, the AIM markers were better able to distinguish populations used in the ssBLUP models.</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

Using genomics to trace genealogy in the early stages of forest tree domestication

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月26日

查看全文
  • Jaroslav Klápště,
  • Natalie J. Graham,
  • Heidi S. Dungey,
  • Mari Suontama,
  • Glenn T. Howe

摘要

Because climate change is pressuring every terrestrial ecosystem including forests, efficient methods for managing forest genetic resources are needed. One approach is to use molecular genetic markers in tree breeding programmes. Current genotyping tools can be used to improve the selection of genetically superior trees via genomic prediction. However, population structure may affect the prediction of breeding values, particularly in the early stages of tree domestication. We examined how models of population structure affected breeding values using a single-step genomic prediction framework (ssBLUP) in Douglas-fir (Pseudotsuga menziesii (Mirbel) Franco) in New Zealand. In particular, we tested three metafounder scenarios. The ssBLUP1 scenario assumed there was a single metapopulation. The other scenarios assumed there were multiple metapopulations with either no gene flow (ssBLUP2) or gene flow among populations (ssBLUP3). All three metafounder scenarios performed better than the pedigree-based (ABLUP) and benchmark (ssBLUP) scenarios. The multiple metapopulation scenarios (ssBLUP2 and ssBLUP3) had slightly greater heritabilities than did the ABLUP and ssBLUP approaches, but heritabilities doubled for the single metapopulation scenario (ssBLUP1). The ssBLUP1 scenario also had the largest accuracy of genomic estimated breeding values (GEBV) but also the largest GEBV bias. In contrast, ssBLUP3 had the lowest GEBV accuracy and bias. Results for the ssBLUP2 scenario demonstrated the value of accounting for genetic variation among metafounders. Compared to the single metapopulation scenario (ssBLUP1), ssBLUP2 had lower GEBV bias and smaller underdispersion of GEBV. Finally, our study demonstrated the benefits of selecting ancestry-informative markers (AIM) for use in the ssBLUP model with metafounders. Compared to all markers, the AIM markers were better able to distinguish populations used in the ssBLUP models.