DeHaan Lee R, Van Tassel David L
The Land Institute, 2440 E. Water Well Rd., Salina, Kansas 67401 USA.
Am J Bot. 2014 Oct;101(10):1801-19. doi: 10.3732/ajb.1400084. Epub 2014 Oct 8.
Annual grain crops dominate agricultural landscapes and provide the majority of calories consumed by humanity. Perennial grain crops could potentially ameliorate the land degradation and off-site impacts associated with annual grain cropping. However, herbaceous perennial plants with constitutively high allocation to harvestable seeds are rare to absent in nature. Recent trade-off theory models suggest that rugged fitness landscapes may explain the absence of this form better than sink competition models. Artificial selection for both grain production and multiyear lifespan can lead to more rapid progress in the face of fitness and genetic trade-offs than natural selection but is likely to result in plant types that differ substantially from all current domestic crops. Perennial grain domestication is also likely to require the development of selection strategies that differ from published crop breeding methods, despite their success in improving long-domesticated crops; for this purpose, we have reviewed literature in the areas of population and evolutionary genetics, domestication, and molecular biology. Rapid domestication will likely require genes with large effect that are expected to exhibit strong pleiotropy and epistasis. Cryptic genetic variation will need to be deliberately exposed both to purge mildly deleterious alleles and to generate novel agronomic phenotypes. We predict that perennial grain domestication programs will benefit from population subdivision followed by selection for simple traits in each subpopulation, the evaluation of very large populations, high selection intensity, rapid cycling through generations, and heterosis. The latter may be particularly beneficial in the development of varieties with stable yield and tolerance to crowding.
一年生谷物作物主导着农业景观，并提供了人类消耗的大部分热量。多年生谷物作物有可能改善与一年生谷物种植相关的土地退化和场外影响。然而，自然中很少存在或根本不存在将大量资源持续分配到可收获种子上的草本多年生植物。最近的权衡理论模型表明，崎岖的适应度景观可能比汇竞争模型能更好地解释这种形式的缺失。与自然选择相比，对谷物产量和多年寿命进行人工选择在面对适应度和基因权衡时能够带来更快的进展，但可能会产生与所有当前驯化作物有很大差异的植物类型。多年生谷物驯化可能还需要制定与已发表的作物育种方法不同的选择策略，尽管这些方法在改良长期驯化的作物方面取得了成功；为此，我们回顾了群体与进化遗传学、驯化和分子生物学领域的文献。快速驯化可能需要具有大效应的基因，这些基因预计会表现出强烈的多效性和上位性。需要刻意暴露隐藏的遗传变异，以清除轻度有害的等位基因并产生新的农艺表型。我们预测，多年生谷物驯化计划将受益于群体细分，随后在每个亚群体中选择简单性状、评估非常大的群体、高选择强度、快速世代循环以及杂种优势。后者在开发产量稳定且耐密植的品种时可能特别有益。
