Conflicts of CpG density and DNA methylation are proximally and distally involved in gene regulation in human and mouse tissues
Conflicts of CpG density and DNA methylation are proximally and distally involved in gene regulation in human and mouse tissues
复制标题
CpG 密度和 DNA 甲基化的冲突在近端和远端参与人类和小鼠组织的基因调控
DOI：
10.1080/15592294.2018.1500057
复制
发表时间：
2018-01-01
期刊：
影响因子：
3.7
通讯作者：
Li, Zhiguang
中科院分区：
文献类型：
--
作者：
Chen, Fushun;Zhang, Qingzheng;Li, Zhiguang
摘要近年来，CpG含量与DNA甲基化的关系引起了人们的极大兴趣。直接或间接的方法已被开发，以调查其调节功能的基础上各种假设，大型队列研究和荟萃分析。然而，所有这些分析都是以CpG块为单位进行的，因此忽略了更精细的基因组结构的影响。在此，我们提出了一种新的碱基对解析算法，系统地研究了CpG含量和DNA甲基化之间的关系。通过引入“互补指数”的概念，我们检测了34个成人和7个胚胎组织的甲基化，并成功地将DNA甲基化和CpG密度的关系拟合到一个非线性数学模型中。开发了另一种算法来定位CpG密度与模型预期不匹配的区域，称为“间隙冲突”（COG）区域。有趣的是，COG在人类和小鼠中是高度一致的，并且它们的分布显示出组织特异性模式。基于COG甲基化模式，我们根据组织的功能或来源对它们进行了正确的分类。我们证明，基于我们的方法的COG可以揭示更多和更深的信息比传统的差异甲基化区域（DMR）的方法。我们还发现，当COG位于转录起始位点（TSS）附近时，这些区域可以决定哪些启动子将用于启动基因转录。此外，位于远离TSS的COG在组蛋白修饰、序列保守、转录因子结合和DNA酶I超敏性方面作为增强子。
ABSTRACT The relationship between CpG content and DNA methylation has attracted considerable interest in recent years. Direct or indirect methods have been developed to investigate their regulatory functions based on various hypotheses, large cohort studies, and meta-analyses. However, all of these analyses were performed at units of CpG blocks and, thus, the influence of finer genome structure has been neglected. Herein, we present a novel algorithm of base-pair resolution to systematically investigate the relationship between CpG contents and DNA methylation. By introducing the concept of ‘complementary index’ we examined the methylomes of 34 adult and 7 embryonic tissues and successfully fitted the relationship of DNA methylation and CpG density into a nonlinear mathematical model. A further algorithm was developed to locate the regions where CpG density does not match expectations from the model, termed ‘conflict of gap’ (COG) regions. Interestingly, COGs are highly concordant in human and mouse and their distributions display a tissue-specific pattern. Based on COG methylation patterns we correctly classified tissues according to their function or origin. We demonstrate that COGs based on our method can reveal more and deeper information than traditional differential methylation region (DMR) approaches. We also found that when COGs are located near to transcription start site (TSS), these regions can determine which promoters will be utilized for initiating gene transcription. Furthermore, COGs located far from the TSS perform as enhancers in terms of histone modification, sequence conservation, transcription factor binding, and DNase I-hypersensitivity.