SM-RCNV: a statistical method to detect recurrent copy number variations in sequenced samples
SM-RCNV: a statistical method to detect recurrent copy number variations in sequenced samples
复制标题
SM-RCNV：一种检测测序样本中反复出现的拷贝数变异的统计方法
DOI：
10.1007/s13258-019-00788-9
复制
发表时间：
2019-05-01
期刊：
影响因子：
2.1
通讯作者：
Jiang, Shan
中科院分区：
文献类型：
--
作者：
Li, Yaoyao;Yuan, Xiguo;Jiang, Shan
拷贝数变异(Copy Number Variation，CNV)是基因组结构变异的一种重要形式，与人类多种疾病有关。利用下一代测序(NGS)数据并开发计算方法来表征这些结构变异对于了解疾病的机制具有重要意义。目的本研究的目的是开发一种新的统计方法来检测来自基因组序列的多个样本中的复发CNV。该方法通过组合整个样本在一个位置的变异频率和连续位置之间的相关性来使用与每个位置相关联的统计量。使用具有已知CNV的真实数据集来训练频率和相关性的权重。结果在接收者工作特性曲线下，SM-RCNV与6种PEER方法相比具有更好的性能。SM-RCNV成功地识别了许多一致的复发CNV，其中大多数已知具有生物学意义并与疾病基因相关。SM-RCNV在CEU呼叫集和YRI呼叫集的符合率分别为258/328(79%)和(157/309)51%。结论SM-RCNV是一种从多个基因组序列中检测复发CNV的统计框架，为人类疾病的基因组研究提供了有价值的信息。源代码可以在https://sourceforge.net/projects/sm-rcnv/上免费获得。
BackgroundCopy number variation (CNV) is an important form of genomic structural variation and is linked to dozens of human diseases. Using next-generation sequencing (NGS) data and developing computational methods to characterize such structural variants is significant for understanding the mechanisms of diseases.ObjectiveThe objective of this study is to develop a new statistical method of detection recurrent CNVs across multiple samples from genomic sequences.MethodsA statistical method is carried out to detect recurrent CNVs, referred to as SM-RCNV. This method uses a statistic associated with each location by combining the frequency of variation at one location across whole samples and the correlation among consecutive locations. The weights of the frequency and correlation are trained using real datasets with known CNVs. P-value is assessed for each location on the genome by permutation testing.ResultsCompared with six peer methods, SM-RCNV outperforms the peer methods under receiver operating characteristic curves. SM-RCNV successfully identifies many consistent recurrent CNVs, most of which are known to be of biological significance and associated with diseased genes. The validation rate of SM-RCNV in the CEU call set and YRI call set with Database of Genomic Variants are 258/328 (79%) and (157/309) 51%, respectively.ConclusionSM-RCNV is a well-grounded statistical framework for detecting recurrent CNVs from multiple genomic sequences, providing valuable information to study genomes in human diseases. The source code is freely available at https://sourceforge.net/projects/sm-rcnv/ .