Assessing the performance of the haplotype block model of linkage disequilibrium
Assessing the performance of the haplotype block model of linkage disequilibrium
复制标题
DOI：
10.1086/378099
复制
发表时间：
2003-09-01
影响因子：
9.8
通讯作者：
Pritchard, JK
中科院分区：
文献类型：
--
作者：
Wall, JD;Pritchard, JK
最近的几项研究表明，人类基因组中的连锁不平衡（LD）具有基本的“块状”结构。然而，到目前为止，很少有正式的评估，以及单倍型块模型捕捉LD的潜在结构。在这里，我们提出了定量的标准来评估如何块状LD是和应用这些标准的真实的和模拟数据。几个大的数据集的分析表明，真实的数据显示出部分适合的单倍型块模型，一些地区符合得很好，而其他人没有。一些改进可以通过基因分型更高的标记密度，但不是通过增加样本的数量。尽管如此，虽然真实的数据只是适度的块状，我们的模拟表明，在均匀重组的模型下，LD的结构实际上不太适合块模型。模拟的模型，其中大部分的重组发生在狭窄的热点提供了一个更好的适合所观察到的模式LD，这表明有广泛的精细尺度的变化，在整个人类基因组的重组率。
Several recent studies have suggested that linkage disequilibrium (LD) in the human genome has a fundamentally "blocklike" structure. However, thus far there has been little formal assessment of how well the haplotype block model captures the underlying structure of LD. Here we propose quantitative criteria for assessing how blocklike LD is and apply these criteria to both real and simulated data. Analyses of several large data sets indicate that real data show a partial fit to the haplotype block model; some regions conform quite well, whereas others do not. Some improvement could be obtained by genotyping higher marker densities but not by increasing the number of samples. Nonetheless, although the real data are only moderately blocklike, our simulations indicate that, under a model of uniform recombination, the structure of LD would actually fit the block model much less well. Simulations of a model in which much of the recombination occurs in narrow hotspots provide a much better fit to the observed patterns of LD, suggesting that there is extensive fine-scale variation in recombination rates across the human genome.