Payseur Bret A, Cutter Asher D, Nachman Michael W
Department of Ecology and Evolutionary Biology, Biosciences West Building, University of Arizona, Tucson, Arizona 85721, USA.
Mol Biol Evol. 2002 Jul;19(7):1143-53. doi: 10.1093/oxfordjournals.molbev.a004172.
Both natural selection and nonequilibrium population-level processes can lead to a skew in the frequency distribution of polymorphisms. Population-level processes are expected to affect all loci in a roughly equal fashion, whereas selection will affect only some regions of the genome. We conducted a sliding-window analysis of the frequency distribution of microsatellite polymorphisms across the human genome to identify regions that may be under positive selection. The analysis was based on a published data set of 5,257 mapped microsatellites in individuals of European ancestry. Observed and expected numbers of alleles were compared under a stepwise mutation model (SMM) using analytical formulae. Observed and expected heterozygosities were compared under a SMM using coalescent simulations. The two sets of analyses gave similar results. Approximately one-fourth of all loci showed a significant deficit of heterozygosity, consistent with a recent population expansion. Forty-three windows were identified with extreme skews in the frequency distribution of polymorphisms (in the direction of a deficit of heterozygosity, given the number of alleles). If these extreme windows are tracking selection at linked sites, theory predicts that they should be more common in regions of the genome with less recombination. We tested this prediction by comparing recombination rates in these extreme windows and in other regions of the genome and found that extreme windows had a significantly lower recombination rate than the genomic average. The proportion of extreme windows was significantly higher on the X chromosome than on the autosomes. Moreover, all the windows with extreme skews on the X chromosome were found in two clusters near the centromere; both these clusters exhibit markedly reduced recombination rates. These analyses point to regions of the genome that may recently have been subject to positive selection. These results also suggest that the effects of positive selection may be more pronounced on the X chromosome than on the autosomes in humans.
自然选择和非平衡种群水平过程都可能导致多态性频率分布的偏差。种群水平过程预计会以大致相同的方式影响所有基因座，而选择只会影响基因组的某些区域。我们对人类基因组中微卫星多态性的频率分布进行了滑动窗口分析，以确定可能受到正选择的区域。该分析基于一个已发表的数据集，该数据集包含5257个在欧洲血统个体中定位的微卫星。使用解析公式在逐步突变模型（SMM）下比较观察到的和预期的等位基因数量。使用合并模拟在SMM下比较观察到的和预期的杂合度。两组分析得出了相似的结果。大约四分之一的所有基因座显示出显著的杂合度不足，这与近期的种群扩张一致。确定了43个窗口，其多态性频率分布存在极端偏差（鉴于等位基因数量，偏向杂合度不足的方向）。如果这些极端窗口追踪的是连锁位点的选择，理论预测它们在基因组中重组较少的区域应该更常见。我们通过比较这些极端窗口和基因组其他区域的重组率来检验这一预测，发现极端窗口的重组率显著低于基因组平均水平。X染色体上极端窗口的比例显著高于常染色体。此外，X染色体上所有具有极端偏差的窗口都位于着丝粒附近的两个簇中；这两个簇的重组率都明显降低。这些分析指出了基因组中可能最近受到正选择的区域。这些结果还表明，在人类中，正选择对X染色体的影响可能比对常染色体更为明显。
