应用简化基因组测序（GBS)与全基因组选择(GS)提高马尾松育种效率的理论性探讨

The breakthrough in next-generation sequencing (NGS) has been accelerating the development of new breeding techniques on genotyping-by-sequencing (GBS) and Genomic selection (GS). GS can perform early selection according to individual genotype because GS estimates the genomic estimated breeding values (GEBVs) through the application of GS model by consummating all effects of genes across whole genome. GS has become an increasing hot topic in forest tree breeding, thanks to its potential of revolutionizing tree breeding by greatly shortening tree breeding cycle, and then it has become a suitable choice in promoting forest tree breeding efficiency. However, GS has not been fully explored for Chinese forest tree breeding. The difficult on high-throughput molecular marker discovery in trees, the complexity of computing and testing in GS model application and the vague understanding on GS hinder the application of GS in Chinese forestry. Fewer Chinese researchers have understood how GS works and even questioned whether GS could really improve the tree breeding efficiency. GBS has implemented innovational DNA marker discovery in high-throughput SNPs. This proposed research will focus on the theoretical investigation of Masson pine (Pinus massoniana Lamb.) breeding efficiency through GBS and GS. First, we will develop a high-throughput GBS SNPs (single nucleotide polymorphisms) discovery platform in Masson pine. Second, we will perform computer simulations to assess the effectiveness of various GS models under various genetic and breeding scenarios. Third, we will validate the assessed GS models with the empirical data collected from Masson pine breeding populations and compare the breeding efficiencies under different GS-assisted Masson pine breeding programs. Fourth, we will develop effective GS-assisted breeding strategies for Masson pine and other forest trees. It is our goal that this research will generate useful information for better understanding of GS-assisted tree breeding, establish a theoretical investigation model on GS-assisted breeding, decrypt the complexity of computing and testing in GS model application, provide a theoretical basis for developing Masson pine breeding strategies, and enhance the GS-assisted tree breeding efforts in China.

下一代测序技术的突破促进了简化基因组测序（GBS）与全基因组选择（GS）育种新技术的快速发展，GS辅助育种依据GS模型估计GEBVs进行基因型选择，这对缩短林木育种周期，提高育种效率潜力巨大。然而，林木高密度分子标记开发困难、GS模型计算与测试的复杂性，以及人们对GS辅助育种的模糊认识和对GS是否真能提高林木育种效率的疑虑，正严重地限制着国内林木GS辅助育种研究的发展。而GBS拥有高通量SNPs开发的潜力，因此，本项目开展应用GBS与GS提高马尾松育种效率的理论性探讨，综合应用多种林木育种新技术，1）研究建立马尾松高通量GBS-SNPs分子标记平台；2）应用计算机模拟，探讨不同GS模型在不同遗传与育种场景中的表现，开展有效GS模型选择与测试，进行提升GS模型预测准确性的理论性探讨；3）开展不同GS方案下马尾松育种效率比较研究；4）发展马尾松及其他林木GS辅助育种研究策略。本项目的实施，期望能为我们提供理解林木GS辅助育种程序的案例，从而建立国内林木GS辅助育种理论性探讨模式，解密GS模型计算与测试的复杂性，加速马尾松GS辅助育种研究策略的发展，为国内林木遗传改良提供理论创新与技术突破。

基于简化基因组测序（GBS）开发马尾松高通量SNP用于马尾松全基因组选择（GS）育种研究可为我国林木GS育种搭建GS新技术育种实施的技术平台。项目对马尾松育种群体363样本在胸径、树高、材积等性状差异上选择取样，分别开展了马尾松GBS测序、SNP挖掘和GS预测准确性的理论和实践探讨，应用计算机模拟理论性探讨了林木GS育种策略，R语言编程开发遗传分析软件AveDissR，马尾松GBS测序数据生物信息学分析SNP开发，撰写Shell和R语言编程指令集流程化处理和调用GS育种的各个环节、开展GS模型调用和马尾松GS育种探讨等，结果表明：实施基于差异性状选择的RNA-Seq测序发展功能性SNP的林木GS育种策略，理论上能极大地提升GS预测准确性；马尾松GBS测序分析获得了577万多个SNP分子标记；新开发出的遗传分析软件AveDissR，既简化了相关遗传分析参数估计的繁琐流程，也提升了遗传分析软件对高通量分子标记数据的处理能力；把多种生物信息学软件分析功能模块调用、多线程并行处理与计算机脚本编程优化相结合，搭建了一个针对林木多样本、高通量GBS测序数据SNP挖掘的高效便捷高通量SNP分子标记开发应用技术平台；不同分子标记类型GS育种计算机模拟探讨的结果揭示了标记类型对GS育种预测准确性提升的影响趋势；基于功能模块编程流程化处理GS育种各个环节的编程代码指令集，搭建起林木GS育种应用技术平台，解密了林木GS育种的复杂性；实施马尾松高通量SNP应用于GS育种的理论性探讨结果表明，应用全基因组分布的随机性SNP其GS预测准确率仅在17-33%之间，而基于理想状态下的QTL-SNP的GS育种其理论预测准确率可高达64-86%。本项目的实施，从理论和实践两方面解密了林木GS育种的流程、技术要求及其复杂性，研究结果为我国林木GS育种的应用和推广奠定了理论指导和关键技术支撑。