Table_3_Boosting Genetic Gain in Allogamous Crops via Speed Breeding and Genomic Selection.xlsx

Breeding schemes that utilize modern breeding methods like genomic selection (GS) and speed breeding (SB) have the potential to accelerate genetic gain for different crops. We investigated through stochastic computer simulation the advantages and disadvantages of adopting both GS and SB (SpeedGS) into commercial breeding programs for allogamous crops. In addition, we studied the effect of omitting one or two selection stages from the conventional phenotypic scheme on GS accuracy, genetic gain, and inbreeding. As an example, we simulated GS and SB for five traits (heading date, forage yield, seed yield, persistency, and quality) with different genetic architectures and heritabilities (0.7, 0.3, 0.4, 0.1, and 0.3; respectively) for a tall fescue breeding program. We developed a new method to simulate correlated traits with complex architectures of which effects can be sampled from multiple distributions, e.g. to simulate the presence of both minor and major genes. The phenotypic selection scheme required 11 years, while the proposed SpeedGS schemes required four to nine years per cycle. Generally, SpeedGS schemes resulted in higher genetic gain per year for all traits especially for traits with low heritability such as persistency. Our results showed that running more SB rounds resulted in higher genetic gain per cycle when compared to phenotypic or GS only schemes and this increase was more pronounced per year when cycle time was shortened by omitting cycle stages. While GS accuracy declined with additional SB rounds, the decline was less in round three than in round two, and it stabilized after the fourth SB round. However, more SB rounds resulted in higher inbreeding rate, which could limit long-term genetic gain. The inbreeding rate was reduced by approximately 30% when generating the initial population for each cycle through random crosses instead of generating half-sib families. Our study demonstrated a large potential for additional genetic gain from combining GS and SB. Nevertheless, methods to mitigate inbreeding should be considered for optimal utilization of these highly accelerated breeding programs.

采用基因组选择（genomic selection, GS）与快速育种（speed breeding, SB）等现代育种手段的育种方案，有望加快多种作物的遗传增益速率。我们通过随机计算机模拟，探究了将GS与SB结合（即SpeedGS）应用于异花授粉作物商业化育种项目的利弊。此外，我们还研究了从常规表型选择方案中省略1或2个选择阶段，对基因组选择精度、遗传增益以及近交程度的影响。以高羊茅育种项目为例，我们针对5个性状——抽穗期、牧草产量、种子产量、持久性与品质——进行了GS与SB的模拟，这些性状具有不同的遗传架构和遗传力（分别为0.7、0.3、0.4、0.1和0.3）。我们开发了一种新方法，可模拟具有复杂遗传架构的相关性状：其效应可从多种分布中抽样，例如用于模拟同时存在微效基因与主效基因的情况。常规表型选择方案的育种周期为11年，而本文提出的SpeedGS方案的单周期时长为4至9年。总体而言，SpeedGS方案在所有性状上均实现了更高的年遗传增益，对于持久性这类低遗传力性状尤为显著。研究结果表明，与仅采用表型选择或仅采用GS的方案相比，增加SB轮次可提升单周期遗传增益；而通过省略育种阶段缩短周期时长后，年遗传增益的提升效果更为显著。尽管GS精度会随SB轮次增加而下降，但第3轮的精度降幅小于第2轮，且在第4轮SB后精度趋于稳定。不过，SB轮次增加会提升近交速率，这可能会限制长期遗传增益的提升。若通过随机杂交而非半同胞家系来构建每个周期的初始群体，近交速率可降低约30%。本研究证实，将GS与SB结合可大幅提升遗传增益潜力。不过，为实现这类高倍速育种方案的最优利用，需考虑采用缓解近交问题的相关方法。