To assess the potential of detecting copy number variations (CNVs) directly from exome sequencing (ES) data in diagnostic settings, we developed a CNV-detection pipeline based on ExomeDepth software and applied it to ES data of 450 individuals. Initially, only CNVs affecting genes in the requested diagnostic gene panels were scored and tested against arrayCGH results. Pathogenic CNVs were detected in 18 individuals. Most detected CNVs were larger than 400 kb (11/18), but three individuals had small CNVs impacting one or a few exons only and were thus not detectable by arrayCGH. Conversely, two pathogenic CNVs were initially missed, as they impacted genes not included in the original gene panel analysed, and a third one was missed as it was in a poorly covered region. The overall combined diagnostic rate (SNVs + CNVs) in our cohort was 36%, with wide differences between clinical domains. We conclude that (1) the ES-based CNV pipeline detects efficiently large and small pathogenic CNVs, (2) the detection of CNV relies on uniformity of sequencing and good coverage, and (3) in patients who remain unsolved by the gene panel analysis, CNV analysis should be extended to all captured genes, as diagnostically relevant CNVs may occur everywhere in the genome.

中文翻译：

---

罕见遗传疾病常规诊断中外显子组测序数据的 CNV 检测：机遇和局限

为了评估在诊断环境中直接从外显子组测序 (ES) 数据检测拷贝数变异 (CNV) 的潜力，我们开发了一个基于 ExomeDepth 软件的 CNV 检测管道，并将其应用于 450 个人的 ES 数据。最初，仅对所需诊断基因组中影响基因的 CNV 进行评分并针对 arrayCGH 结果进行测试。在 18 个人中检测到致病性 CNV。大多数检测到的 CNV 大于 400 kb (11/18)，但三个个体的 CNV 仅影响一个或几个外显子，因此无法被 arrayCGH 检测到。相反，最初遗漏了两个致病性 CNV，因为它们影响了未包含在分析的原始基因组中的基因，而第三个被遗漏了，因为它位于覆盖不良的区域。我们队列中的总体综合诊断率（SNVs + CNVs）为 36%，临床领域之间存在很大差异。我们得出结论：（1）基于 ES 的 CNV 管道可有效检测大小致病性 CNV，（2）CNV 的检测依赖于测序的一致性和良好的覆盖率，以及（3）在基因组分析仍未解决的患者中, CNV 分析应扩展到所有捕获的基因，因为诊断相关的 CNV 可能出现在基因组的任何地方。