一种碱基精度的肿瘤基因组单体型异质性识别算法

发布时间：2019-01-08 16:41

【摘要】：针对肿瘤组织的异质性的子克隆解析,提出了一种通过多级子克隆的体细胞突变模式来识别单体型异质性的算法。该算法基于肿瘤组织的多文库测序数据提取文库特征和双末端读段约束,通过对体细胞突变位点的等位基因变异频率进行聚类估算出子克隆数目的一个先验;同时设计了一种拼接识别算法,通过遍历位点对应的读段来拼接单体型序列,拼接出的单体型序列的精度为碱基水平;采用后验概率的最大似然估计解出子克隆的个数、配比及演化关系。仿真实验表明,当基础文库满足一定测序覆盖度时,该算法对单体型异质性的识别精度可达到99%以上,能够取代目前数据分析中常用的两步法,且获得高精确的识别结果。
[Abstract]:Based on the analysis of the heterogeneity of tumor tissues, an algorithm is proposed to identify haplotype heterogeneity by the somatic mutation pattern of multistage cloning. Based on the multilibrary sequencing data of tumor tissues, the algorithm extracts library features and double-terminal reading constraints, and estimates a priori number of clones by clustering the allelic mutation frequency of somatic mutation sites. At the same time, a splicing recognition algorithm is designed, which splicing haplotype sequences by traversing the corresponding reading segments of the site, and the accuracy of the stitched haplotype sequences is base level. The maximum likelihood estimation of the posterior probability is used to calculate the number, ratio and evolution of the subclones. The simulation results show that the recognition accuracy of the algorithm for haplotype heterogeneity can reach more than 99% when the base library meets a certain sequencing coverage, which can replace the two-step method commonly used in data analysis and obtain high accurate recognition results.
【作者单位】：锦州医科大学公共基础学院;西安交通大学计算机科学与技术系;西安交通大学陕西省医疗健康大数据工程研究中心;第四军医大学肿瘤生物学国家重点实验室;西安交通大学管理学院;
【基金】：国家自然科学基金资助项目(81400632) 陕西省自然科学基金资助项目(2014JM8350) 中央高校基本科研业务费专项资金资助项目(GLIJ002)
【分类号】：R73-3

【相似文献】