Obtaining informative data is the ambition of any genomic project, but in nonmodel species with very large genomes, pursuing such a goal requires surmounting a series of analytical challenges. Double‐digest RAD sequencing is routinely used in nonmodel organisms and offers some control over the volume of data obtained. However, the volume of data recovered is not always an indication of the reliability of data sets, and quality checks are necessary to ensure that true and artefactual information is set apart. In the present study, we aim to fill the gap existing between the known applicability of RAD sequencing methods in plants with large genomes and the use of the retrieved loci for population genetic inference. By analysing two populations of Cypripedium calceolus, a nonmodel orchid species with a large genome size (1C ~ 31.6 Gbp), we provide a complete workflow from library preparation to bioinformatic filtering and inference of genetic diversity and differentiation. We show how filtering strategies to dismiss potentially misleading data need to be explored and adapted to data set‐specific features. Moreover, we suggest that the occurrence of organellar sequences in libraries should not be neglected when planning the experiment and analysing the results. Finally, we explain how, in the absence of prior information about the genome of the species, seeking high standards of quality during library preparation and sequencing can provide an insurance against unpredicted technical or biological constraints.

中文翻译：

---

尽管基因组很大，但有效的双消化 RAD 测序和基因分型

获取信息数据是任何基因组项目的目标，但在具有非常大基因组的非模型物种中，实现这一目标需要克服一系列分析挑战。双消化 RAD 测序通常用于非模式生物，并提供对获得的数据量的一些控制。然而，恢复的数据量并不总是表明数据集的可靠性，质量检查是必要的，以确保将真实信息和人工信息区分开来。在本研究中，我们的目标是填补 RAD 测序方法在具有大基因组的植物中的已知适用性与使用检索到的基因座进行种群遗传推断之间存在的差距。通过分析两个Cypripedium calceolus种群，一种具有大基因组大小 (1C ~ 31.6 Gbp) 的非模型兰花物种，我们提供了从文库制备到生物信息过滤和遗传多样性和分化推断的完整工作流程。我们展示了如何探索过滤策略以消除潜在的误导性数据并使其适应数据集特定的特征。此外，我们建议在计划实验和分析结果时不应忽视文库中细胞器序列的发生。最后，我们解释了在缺乏有关物种基因组的先验信息的情况下，如何在文库制备和测序过程中寻求高标准的质量可以为应对不可预测的技术或生物学限制提供保障。