Kalinowski S T, Hedrick P W
Department of Biology, Arizona State University, Tempe, AZ 85287, USA.
Heredity (Edinb). 2001 Dec;87(Pt 6):698-708. doi: 10.1046/j.1365-2540.2001.00966.x.
The great expansion of population genetic data using molecular techniques now allows examination of the extent of linkage disequilibrium for many pairs of loci, each locus often with multiple alleles. The expectation-maximization (EM) algorithm for generating maximum likelihood estimates of gametic frequencies from multiallelic genotypic data is described and applied. The EM algorithm is used in desert bighorn sheep where the population size, and consequently the sample size, is often small. We calculated haplotype frequencies for all pairwise combinations of five major histocompatibility loci and three microsatellite loci in 14 populations; the performance of the algorithm is discussed. Disequilibrium values are calculated and tested for statistical significance. High levels of disequilibrium are found between all pairs of major histocompatibility complex (MHC) loci and between MHC and a linked microsatellite locus.
利用分子技术对群体遗传数据的大幅扩充，如今使得我们能够对许多基因座对的连锁不平衡程度进行检测，每个基因座通常具有多个等位基因。本文描述并应用了期望最大化（EM）算法，该算法用于从多等位基因基因型数据中生成配子频率的最大似然估计值。EM算法被应用于沙漠大角羊群体，该群体的种群规模以及相应的样本量往往较小。我们计算了14个群体中五个主要组织相容性基因座和三个微卫星基因座所有成对组合的单倍型频率，并讨论了该算法的性能。计算了不平衡值并检验其统计学显著性。结果发现，所有主要组织相容性复合体（MHC）基因座对之间以及MHC与一个连锁微卫星基因座之间都存在高水平的不平衡。
