nana进来~
2.4. Data analysis
Genetic polymorphism was estimated for each popu¬lation as the number of alleles per locus (A), observed heterozygosity (HO) and expected heterozygosity (HE), computed using GENEPOP 3.1 (Raymond andRousset, 1995) and ARLEQUIN 2.000 (Schneider et al., 2000). The probability test of Guo and Thompson (1992) was used to test for genotypic linkage dis¬equilibrium for all pairs of loci within each population and to test for locus conformity to Hardy-Weinberg equilibrium (Raymond and Rousset, 1995). The dis¬tribution of population-specific alleles detected in A. hispanica was analysed following Gibbs et al. (1997).
Due to the large difference in size between the various samples, A, HO and HE were compared between popu¬lations using a Monte-Carlo simulation, re-sampling 1000 times from each of the larger samples with a sam¬ple size equal to that of the smallest sampled popula¬tion, Degebe (15 individuals) using the software POPTOOLS v.2.1 (Hood, 2000).
GENEPOP was also used to compute unbiased esti¬mates of the p-values of the homogeneity tests (Ray¬mond and Rousset, 1995) among all pairs of samples for all loci. The Weir and Cockerham (1984) analogues of Wright’s F statistics, FIS (f) and FST (y) and their sig-
nificance per locus were estimated using FSTAT (Gou¬det, 2000). The extent of genetic differentiation among populations was also quantified by computing pairwise FST values and by performing a global test of differ¬entiation among samples (Raymond and Rousset, 1995) using ARLEQUIN.
The significance of the spatial variation in gene diversity among populations was estimated by per-forming a hierarchical analysis of genetic diversity among populations using the analysis of molecular variance model (AMOVA), as described in Michalakis and Excoffier (1996) and available in ARLEQUIN.
The program MANTEL 2.0 (Liedloff, 1999) was used to test the significance of the relationship between geo¬graphical distance and the FST values for all pairs of populations. Finally, an AMOVA (Michalakis and Excoffier, 1996) was performed using ARLEQUIN with an a priori design which divided all populations into three groups corresponding to the dominant mtDNA lineages previously suggested as ESUs.
参考答案:2.4. 数据分析
估计了每个种群的遗传多态性并用GENEPOP 3.1 (Raymond和Rousset，1995)和ARLEQUIN 2.000 (Schneider等，2000)软件计算，这种遗传多态性包括下列3个指标：每基因座的等位基因数目(A)、观察杂合度(HO)和预期杂合度 (HE)。Guo 和Thompson (1992)的概率测验用来测验每个种群内所有基因座对之间的基因连锁不平衡，并用来测试基因座对哈迪－温伯格平衡的适合度 (Raymond和Rousset，1995)。西班牙鱥中检测到的种群特异性等位基因的分布遵照Gibbs等(1997)的方法分析。
由于不同样本的容量之间差异很大，用蒙特卡罗模拟法比较种群之间的每基因座等位基因数、观察杂合度和预期杂合度，用POPTOOLS v.2.1 (Hood，2000)软件从每个较大样本中重新取样1000次使所取样本的容量和最小种群（Degebe ：15个个体)的样本容量相当。
GENEPOP软件也用来计算所有样本对的所有基因座之间同质性测验 (Raymond和Rousset，1995) 中p-值的无偏估计量。用FSTAT (Goudet，2000)软件估计Wright氏 F 统计的Weir 和Cockerham (1984) 相似序列、每基因座的近交系数(f)和基因分化系数（FST）(y)及其显著性。种间遗传分化程度也用计算配对基因分化系数的方法进行定性，并用ARLEQUIN软件(Raymond和Rousset, 1995)来执行样本之间分化的球状测试。
种间基因多样性空间变异的显著性用在分子变异模型(AMOVA)上操作的种间遗传多样性层次分析来估计，如同Michalakis和Excoffier (1996)描述的，并且在 ARLEQUIN软件中也有效。
用MANTEL 2.0 (Liedloff, 1999)程序来测验所有种群对的地理距离和基因分化系数值之间关系的显著性。最后，把所有种群根据主要线粒体DNA链（以前称之为进化显著性单元）划分成3个群并用带有优先性设计的ARLEQUIN软件构建分子变异模型 (Michalakis和Excoffier, 1996)。