【作者】 苏海林；

【导师】 包振民；

【作者基本信息】 中国海洋大学 ， 遗传学， 2013， 博士

【摘要】 贝类海水养殖是中国的海水养殖的主导产业之一。为了提高贝类生产产量，多性状最佳线性无偏预测(Best Linear Unbiased Prediction, BLUP)选择育种已经被引入到贝类养殖工作中。随着越来越多的遗传和表型信息的积累，如何迅速和有效地将这些信息应用在贝类养殖和育种工作中巳经显得日益重要。基因组选择(Genomic Selection, GS)通过高密度的单核苷酸多态性(Single NucleotidePolymorphism, SNP)的标记计算出基因组育种值，现己被广泛应用于家畜动物和植物物种的育种工作中。最近的研究成果表明，在育种中的应用全基因组选择St遗传改良工作具有革命性的推动作用。本论文分为四个部分：基因组亲缘关系矩阵的构建与真实亲缘关系的计算，全基因组选择软件MixP与gsbay，贝类全基因组选择遗传育种评估与分析平台的构建以及全基因组选择在扇贝育种中的应用。第一部分，提出了一种基于SNP分子标记基因型构建的、适用于衡量水产动物全同胞家系个体之间真实亲缘关系的T矩阵，并通过模拟的数据考察了其与G矩阵、传统的A矩阵在计算个体之间的亲缘相关系数时的差异。虽然T矩阵并不完全适用于估计遗传力与育种值的算法，使用T矩阵在基因组最佳线性无偏预测(Genomic BLUP, GBLUP)中估计育种值的准确性不如G矩阵，但其所估计的两个有亲缘相关关系个体之间的相关系数是由基因组范围分子标记的基因型推算出的精确值，其取值范围在(0.41,1.0)左右，能比A矩阵和G矩阵更好的描述相关关系。第二部分，使用模拟数据比较了两种通过高密度的SNP标记计算出基因组估计育种值(Genomic Estimated Breeding Values, GEBV)的全基因组选择算法，即MixP和gsbay，与GBLUP法在计算基因组估计育种值上的准确性。结果表明，MixP和gsbay都能通过分析运算较准确的估计出基因组估计育种值，并且它们估计准确度的期望均高于GBLUP方法。gsbay比MixP的准确度稍高，而MixP的运算速度比gsbay快，且更适用于影响性状的数量性状基因座(Quantitative Trait Loci, QTL)个数较少时的情况。第三部分，论述了加入GBLUP运算模块到自丨:丌发贝类遗传育种分析评估系统的方案，完成了调用DMU软件进行GBLUP分析功能的嵌入，并整合了扇贝种质资源数据库管理系统，记录和管理贝类的生长性状、繁殖性状、抗逆性状的测量数据以及群体特征、养殖环境和遗传信息；并利用这些信息通过上述模块进行贝类选择育种、遗传性状解析，并能计算遗传力、表型相关、近交系数和育种值，制定选种配种方案以指导选种。第四部分，使用了98个具有基因型信息与表型记录的栉孔扇贝(CWawp全同胞家系真实数据对平台进行了可行性检测，结果表明MixP与gsbay两者均可应用于全基因组选择的实际工作中，在贝类育种工作中展开全基因组选择分析已现实可行。更多还原

【Abstract】 Scallop mariculture is one of the leading industries of sea-farming in China.Aiming at improvement of scallop production, multi-trait selection using best linerunbiased prediction (BLUP) has been introduced to scallop breeding. Withaccumulation of more and more genetic and phenotypic information, there is a needfor rapid and eiffcient utilization of this information to assist in scallop breeding.Genomic selection (GS)? in which selection decisions are made on genomic breedingvalues predicted from high-density single nucleotide polymorphic (SNP) markers，iswidely used in livestock animals and plants species. Recent researches demonstratedthat applying GS in breeding is revolutionizing the genetic improvement.Firstly, this paper proposed a method to construct the T matrix using thegenotypes of SNP molecular markers to assess the real kinship between full-sibindividuals of aquatic animals, and compared it with the C-matrix and the traditionalA-matrix in calculating the correlation coefifcient between individuals withsimulation data. Currently T matrix is not fully applicable to the algorithms thatestimate the heritability and breeding values, and the accuracy of genomic estimatedbreeding value (GEBV) calculated using T-matrix in genomic BLUP (GBLUP) waslower than the G-matrix one, but it could estimate the real correlation coefifcientbetween related individuals with genome-wide molecular markers in the range of(0.41,1.0), which could be a better description than the A-matrix one and theG-matrix one.Subsequently, this study considered two methods, MixP and gsbay，to carry outgenomic selection (GS) analysis using simulated data. The analysis resultsdemonstrated excellent performance and showed that both methods are able toobtain more accurate predictions of GEBV and marker effects in comparison withtraditional genomic selection methods based on GBLUP. The main advantage ofMixP and gsbay is to assume the prior distribution of single nucleotidepolymorphism (SNP) marker effects to be a mixture of two normal distributions inorder to improve accuracy and accelerate the analysis for the high-dimensional data.The study then upgraded the Analysis and Evaluation System for ShellfishGenetic Breeding by equipping it with GBLUP computing modules. And a new system, the Genomic Analysis and Evaluation System tor Shelltish Genetic Breeding, was build to apply genomic selection by calling MixP and gsbay.Afterwards, this study utilized the real data of a full-sib family with98Zhikong scallop (Chlamys farreri) individuals, with both genotype and phenotype records, to test the feasibility of the GS platform. The results indicated that gsbay is more robust, while MixP computes much faster, and both MixP and gsbay can be applied to the actual work of genomic selection:it is feasible to carry out GS in scallop breeding.更多还原