缺失数据下遗传重组率估计的EM方法

M Algorithm-based Method for the Estimation of Genetic Recombination Fractionwith Missing Data

EM算法是在不完全信息资料下实现参数估计的一种通用迭代方法,其在现代科学的许多领域已有着广泛的应用。文章导出了双位点不同标记类型,包括共显性—共显性,共显性—显性和显性—显性三种模式下,部分个体缺失标记基因型时,重组率估计率的EM算法。用编制的SAS/IML程序进行了Monte Carlo模拟研究,验证了文章所述方法在遗传连锁分析中的有效性和实用性。

EM algorithm is a common alternative manner used to implement parameter estimation with incomplete data, which has widely been applied in various fields of modem science. The paper induces the EM algorithm method for generic recombination fraction with parts of individuals missing marker genotypes under different marker types of double locus, three modes in all, including codominance- codominance, codominance- dominance and dominance- dominance. Monte Carlo simulation test was carried out through SAS/IML program in order to verify the validity and utility of the above method in genetic linkage analysis.