Data from: Is local selection so widespread in river organisms? Fractal geometry of river networks leads to high bias in outlier detection

Identifying local adaptation is crucial in conservation biology in order to define ecotypes and establish management guidelines. Local adaptation is often inferred from the detection of loci showing a high differentiation between populations, the so-called FST outliers. Methods of detection of loci under selection are reputed to be robust in most spatial population models. However, using simulations we showed that FST outlier tests provided a high rate of false positives (up to 60%) in fractal environments such as river networks. Surprisingly, the number of sampled demes was correlated with parameters of population genetic structure, such as the variance of FSTs, and hence strongly influenced the rate of outliers. This unappreciated property of river networks therefore needs to be accounted for in genetic studies on adaptation and conservation of river organisms.在保护生物学领域，识别局部适应性（local adaptation）对于界定生态型并制定管理指南至关重要。局部适应性通常可通过检测种群间分化程度极高的基因座（locus，复数为loci）来推断，即所谓的FST异常值（FST outliers）。学界普遍认为，检测受选择基因座的方法在多数空间种群模型中具备良好的稳健性。然而，本研究通过模拟实验证实，在河网这类分形环境（fractal environment）中，FST异常值检验会产生极高的假阳性率（最高可达60%）。令人意外的是，采样居群（sampled deme）的数量与种群遗传结构参数（如FST的方差）存在相关性，因此会对异常值检出率产生显著影响。因此，河网的这一未被充分认知的特性，在河生生物的适应性研究与保护遗传学研究中必须予以考量。