【作者】 孙华；

【导师】 杨润清；

【作者基本信息】 东北农业大学 ， 动物遗传育种与繁殖， 2003， 硕士

【摘要】 把表型值随着时间（生命时期，年龄，胎次等）或其他可以量化的因素（生理状态，生产水平，代谢率和环境条件等）变化的性状称为动态性状。受动物遗传育种中用来估计动态性状育种值的随机回归测定日模型思想的启发，将关于时间（测定日期）的Legendre多项式镶嵌在遗传模型的每个遗传效应中，以刻画QTL对动态性状变化过程的作用，从而建立起动态性状基因定位的数学模型。 以F₂遗传设计群体为例，论述了动态性状基因定位分析的基本原理，推导了QTL效应大小和位置的极大似然法估计过程，采用Monte Carlo方法模拟产生标记基因型和个体表型值，设置抽样群体大小、测定日密度和QTL遗传力三个因素各三个水平，研究了抽样群体大小和测定日密度对不同遗传力QTL的检测效力的影响，模拟试验采用正交设计安排因素组合。模拟试验结果表明：个体数和测定日密度对动态性状QTL的分析和检测具有几乎相同的作用，而且在相同样本含量条件下两者呈现互补的关系；高遗传力的要比低遗传力的QTL在检测时需要较少的群体和测定日抽样，但不论QTL的遗传力多大，300以上的群体大小和5％以上的测定日密度都可以保证足够高的检测效力。 同时，在假设误差方差为齐次的情况下，引用回归析法对动态性状进行定位分析以代替似然法。模拟试验结果表明：虽然回归法在各个因素组合条件下对误差方差的估计准确度和精确度都低于似然法，但两种方法对QTL的检测效力差别不大，而且与似然方法相比，回归法表现出明显的计算过程简单、计算速度快等优势。 在QTL定位模型中，嵌入基因效应随过程变化的函数，同以往的单独定位分析和多性状定位分析相比，不仅考虑了动态性状的动态变化规律，而且减少了待估参数，提高了QTL的检测效力和定位的准确度。尤其是用Legendre多项式描述动态性状QTL各种基因效应的作用，既可以发挥多项式在计算上的优势，也确保定位分析模型两端具有相同的几何形状或拟合参数之间的等价性。 鉴于动态性状的广泛性及经济重要性，研究该类性状基因定位分析方法对深入揭示其遗传规律、提高其改良效率具有重要的理论和现实意义。更多还原

【Abstract】 Dynamic trait are those phenotypic value changes with time and other quantifiable factors such as age, parities, physical status, performace level and eviornments etc.. In estimation of breeding value for the dynamic trait, the process with dynamic changes was described by a Legendre polynomial on time (test day) of certain order. On this opinion, the effect of dynamic trait loci could also be explained by inserting a legendre polynomial in each genetic factors, and so the mathematic model for mapping dynamic trait loci was constructed.The objective of this study is to propose a general theoretical framework for embedding a dynamic process in the statistical analysis of QTL mapping for dynamic trait. A maximum likelihood based method, implemented with EM algorithm, was induced to estimate QTL location and effects on dynamic process. The newly approach was demonstrated by Monte Carlo simulation and the effect of population size and test day density on detect power for the QTL with different heritbility and the factor combination in simulation study was designed by orthogonal table. It showed that the individual number and test-day density with almost same effect on QTL analysis and detection, and showed complementation with each other in same sample size. It would need less individual and test-day records with higher heritability in the detection for dynamic trait loci. But the power of detection would be sufficient with individuals above 300 and test-day density over 5% even if in lower heritability.That the detection power of regression analysis in different situation, i.e., individual number, test-day density and heritability, under equal residual variance was also studied and compared with maximum likelihood method. It showed similar detection ability as maximum likelihood method in each combination even regression analysis tended to be lower accuracy and lower precision on estimate of residual variance compared with maximum likelihood. And regression analysis performed significant advantage in computation process and computationspeed.Compared with previous separate analysis and multiple trait mapping, it considered the underlying law in changes of dynamic trait by incorporating a function about the changes of gene effect with dynamic process to QTL mapping model. And because of the reduction of parameter estimated, the detection would be more power and more precision. Especially it can be insured that with same geometric shape in both hand of the mapping or the equivalence infitted parameters by using Legendre polynomial to describe the effect of QTL on the processABSTRACTof dynamic trait.In view of the economical importance of dynamic traits and its extensive existence in features, the study for QTL mapping of these trait would be important in theoretical and practical for revealing its genetic law and increasing its improvement efficiency.更多还原