Inferences about human demography based on multilocus analyses of noncoding sequences.
Inferences about human demography based on multilocus analyses of noncoding sequences.
复制标题
基于非编码序列的多位点分析对人类人口统计学的推论。
DOI：
10.1093/genetics/161.3.1209
复制
发表时间：
2002
期刊：
影响因子：
3.3
通讯作者：
Hudson,RichardR
中科院分区：
文献类型：
--
作者：
Pluzhnikov,Anna;DiRienzo,Anna;Hudson,RichardR
来自10个非连锁常染色体非编码区的数据（对来自三个群体中的每个群体的15个个体进行重新测序）用于多位点分析以测试人类人口统计模型。10个区域中的每一个由~2500 bp组成。本文用多位点分析方法，以Tajima的D和Fu及Li的D ~* 的平均值和方差为统计量，对一类近期种群扩展的模型进行了检验。非洲样本（喀麦隆的豪萨人）与恒定人口规模模型和最近扩展的一系列模型兼容。对于这个总体样本，我们估计的置信度集显示了与增长相容的参数值的有限范围。对于一个指数增长率低至1 × 10−3/代，人口增长不太可能在5万年前开始。对于更高的增长率，增长的开始必须是最近的。根据Tajima'sD的平均值，我们从意大利人口的样本被发现是不符合一个恒定的人口规模模型或任何简单的扩展模型。在中国样本中，Tajima的D方差太大，不符合常数种群大小模型或任何简单的扩展模型。
Data from 10 unlinked autosomal noncoding regions, resequenced in 15 individuals from each of three populations, were used in a multilocus analysis to test models of human demography. Each of the 10 regions consisted of ~2500 bp. The multilocus analysis, based on summary statistics (average and variance of Tajima'sDand Fu and Li'sD*), was used to test a family of models with recent population expansion. The African sample (Hausa of Cameroon) is compatible with a constant population size model and a range of models with recent expansion. For this population sample, we estimated confidence sets that showed the limited range of parameter values compatible with growth. For an exponential growth rate as low as 1 × 10−3/generation, population growth is unlikely to have started prior to 50,000 years ago. For higher growth rates, the onset of growth must be more recent. On the basis of the average value of Tajima'sD, our sample from an Italian population was found to be incompatible with a constant population size model or any simple expansion model. In the Chinese sample, the variance of Tajima'sDwas too large to be compatible with the constant population size model or any simple expansion model.