Du Ruchao, Dong Jinxin, Jiang Hua, Qi Minyong, Zhao Zuyao
School of Computer Science and Technology, Liaocheng University, No. 34 Wenhua Road, Liaocheng, 252000, Shandong, China.
Orthopedics Department, Liaocheng People's Hospital, Liaocheng, China.
Sci Rep. 2025 Jul 1;15(1):22145. doi: 10.1038/s41598-025-06527-3.
Copy number variation (CNV) plays an important role in disease susceptibility as a type of intermediate-scale structural variation (SV). Accurate CNV detection is crucial for understanding human genetic diversity, elucidating disease mechanisms, and advancing cancer genomics. A variety of CNV detection tools based on short sequencing reads from next-generation sequencing (NGS) have been developed. Although many researchers have conducted extensive comparisons of the detection performance of various tools, these studies have not fully considered the comprehensive impact of factors such as variant length, sequencing depth, tumor purity, and CNV types on tools performance. Therefore, we selected 12 widely used and representative detection tools to comprehensively compare their performance on both simulated and real data. For the simulated data, we compared their performance across six variant types under 36 configurations, including three variant lengths, four sequencing depths, and three tumor purities. For the real data, we used the overlapping density score (ODS) to evaluate the performance of the 12 detection tools. Additionally, we compared their time and space complexities. In this study, we analyzed the impact of each configuration on the tools and recommended the most suitable detection tools for each scenario. This study provides important guidance for researchers in selecting the appropriate variant detection tools for complex situations.
作为一种中等规模的结构变异（SV），拷贝数变异（CNV）在疾病易感性中起着重要作用。准确检测CNV对于理解人类遗传多样性、阐明疾病机制以及推动癌症基因组学发展至关重要。基于下一代测序（NGS）的短测序读段，已经开发了多种CNV检测工具。尽管许多研究人员对各种工具的检测性能进行了广泛比较，但这些研究并未充分考虑变异长度、测序深度、肿瘤纯度和CNV类型等因素对工具性能的综合影响。因此，我们选择了12种广泛使用且具有代表性的检测工具，在模拟数据和真实数据上全面比较它们的性能。对于模拟数据，我们在36种配置下比较了它们在六种变异类型上的性能，包括三种变异长度、四种测序深度和三种肿瘤纯度。对于真实数据，我们使用重叠密度分数（ODS）来评估这12种检测工具的性能。此外，我们还比较了它们的时间和空间复杂度。在本研究中，我们分析了每种配置对工具的影响，并针对每种情况推荐了最合适的检测工具。这项研究为研究人员在复杂情况下选择合适的变异检测工具提供了重要指导。
