作物数量性状基因组选择新方法及软件研究

Most crop breeding target traits, such as yield and quality, are quantitative traits. Their genetic mechanism is complicated and controlled by a set of interacted gene network, as well as modified by environment factors. Genetic improvement of quantitative traits needs longer selection cycle and generally exhibits low efficiency. How to improve quantitative traits efficiently and increase selection response in unit time through genetic dissection of gene effects is always the frontier research hot in the crop genetics and breeding which needs to solve emergently. With the fast development of high-throughput technology of omics, genome selection (GS) has become the most promising new breeding theory and method. However, there are still many disadvantages for the existing GS methods. For example, no enough attention is payed on the differences in selection method and efficiency between major and minor genes and the effective selection on dominance and epistasis and their interactions with environments, as well as the influence on the estimation of breeding values and the accuracy of genomic selection because of different allelic frequencies. This project will use an immortalized F2 population of Oryza sativa L. as a worked example to develop new GS model and analytical method based on the orthogonal decomposition model, which could be achieved by mapping and selection on the major-effect genes combined with collective selection on all minor-effect genes by GWAS and estimation of breeding value using kinship matrix. While, an internet service platform for parallelized computation of CPU and GPU will be established. The expected results of this research are very valuable in science for dissection of genetic architecture and improvement of selection efficiency and in breeding practice of genetic improvement of quantitative traits.

作物产量、品质等育种目标性状都是数量性状，遗传机理复杂，受多基因互作网络及环境的协同调控，改良存在选择周期长、效率低等困难。如何通过遗传效应剖析，提高性状单位时间的选择响应及效益，持续有效地改良性状，一直是作物遗传育种亟待解决的前沿科学问题。随高通量组学技术的发展，基因组选择(GS)已成为最具前景的育种新理论与方法，但现有GS方法仍存在较多不足，没有充分利用主效基因与微效多基因选择方法和效率上的差异、对显性、上位性及其与环境互作效应的有效选择，忽略了基因型频率对育种值估算及选择精度的影响等。项目以水稻永久F2群体为实例，关联分析结合基于协方差结构的育种值估算方法，发展基于正交分解模型进行主效基因定位选择结合微效多基因整体选择的GS新模型、新方法，研制CPU与GPU混合并行计算的网络服务平台。研究成果对作物数量性状遗传结构剖析、提高选择效率具有重要的科学意义和育种应用价值。

作物产量、品质等育种目标性状都是数量性状，遗传机理复杂，受多基因互作网络及环境的协同调控。随着高通量组学技术的发展，基因组选择(GS)已成为最具前景的育种新理论与方法，但现有GS方法仍存在不足，如没有充分利用主效基因与微效多基因选择方法和效率上的差异，缺乏对显性、上位性及其与环境互作效应的有效选择等问题，课题组针对这一挑战，提出GS的新模型、新方法，研制CPU与GPU混合并行计算的网络服务平台，提高GS育种的精确性和广泛性，开展的研究及取得的成果主要包括：1）在混合线性模型框架下，提出了主微基因整合基因组选择模型，适用于简单加性、加性-显性-上位性以及加性与环境互作等各类遗传模型，提高了GS模型的预测精度；2）实例分析了烟草重组自交系群体农艺性状和烟气有害成分性状，提高了烟草基因组选择精度，为选育烤烟新品系提供了指导意见；3）发展了联合分析多个复杂性状加性、显性、上位性及其与环境互作的基因定位新方法，提高复杂性状多效性基因的定位功效；4）发展了无监督式的选择信号鉴定方法，提高了检测基因组自然选择事件的效率，丰富了相关性状的基因组选择策略；5）发展了稳健的线性回归方法，有效降低了组学中异常数据对参数估计的影响；6）进行了数量性状分析云平台设计与开发，提供了数量性状分析容器化上云方案。.课题研究已有7篇研究论文在SCI期刊正式刊出，申请获得1项发明专利和1项软件著作权，资助培养了9名硕博研究生及1名博士后。课题研究成果对作物数量性状遗传结构剖析、提高选择效率具有重要的科学意义和育种应用价值，举办全国数量遗传分析与QTX定位研讨班，起到了良好社会效果。.