Systematic comparison of somatic variant calling performance among different sequencing depth and mutation frequency
Systematic comparison of somatic variant calling performance among different sequencing depth and mutation frequency
复制标题
DOI：
10.1038/s41598-020-60559-5
复制
发表时间：
2020-02-26
影响因子：
4.6
通讯作者：
Du, Hongli
中科院分区：
文献类型：
--
作者：
Chen, Zixi;Yuan, Yuchen;Du, Hongli
在过去的十年中，肿瘤的治疗取得了显著的进展，例如靶向治疗的成功临床应用。如今，靶向治疗主要基于突变的检测，而下一代测序（NGS）在相关临床研究中发挥着重要作用。突变频率是肿瘤突变检测中的一个主要问题，增加测序深度是一种广泛使用的提高突变识别性能的方法。因此，有必要评估不同测序深度和突变频率以及突变调用工具的效果。在这项研究中，Strelka 2和Mutect 2工具用于检测30个测序深度和突变频率组合的性能。结果表明，绝大多数样品的精密度保持在95%以上。通常，对于较高的突变频率（>= 20%），测序深度>= 200 X足以识别95%的突变;对于较低的突变频率（= 20%），而Mutect 2在突变频率低于10%时表现更好。此外，Strelka 2平均比Mutect 2快17到22倍。本研究通过对不同测序深度和突变频率的系统性比较，为临床基因组学研究体细胞突变的鉴定提供了一个有用和全面的指导。
In the past decade, treatments for tumors have made remarkable progress, such as the successful clinical application of targeted therapies. Nowadays, targeted therapies are based primarily on the detection of mutations, and next-generation sequencing (NGS) plays an important role in relevant clinical research. The mutation frequency is a major problem in tumor mutation detection and increasing sequencing depth is a widely used method to improve mutation calling performance. Therefore, it is necessary to evaluate the effect of different sequencing depth and mutation frequency as well as mutation calling tools. In this study, Strelka2 and Mutect2 tools were used in detecting the performance of 30 combinations of sequencing depth and mutation frequency. Results showed that the precision rate kept greater than 95% in most of the samples. Generally, for higher mutation frequency (>= 20%), sequencing depth >= 200X is sufficient for calling 95% mutations; for lower mutation frequency (= 20%), while Mutect2 performed better when the mutation frequency was lower than 10%. Besides, Strelka2 was 17 to 22 times faster than Mutect2 on average. Our research will provide a useful and comprehensive guideline for clinical genomic researches on somatic mutation identification through systematic performance comparison among different sequencing depths and mutation frequency.