癌症系统表观基因组学新型生物信息统计算法研究

Complex diseases like cancer represent an enormous economic and social burden to modern society. To improve disease management, requires not only novel biomarkers, but importantly also a better understanding of the causal molecular changes underlying the carcinogenic process. Recent work has shown that many of the novel cancer driver mutations target epigenetic enzymes, thus linking changes in the cancer genome to those in the epigenome. However, how the deregulation of epigenetic enzymes in cancer affects the epigenome and how this in turn disrupts the regulatory networks has not been extensively explored..The main aim of this proposal is to develop bioinformatic methods to perform a pan-cancer-wide systems-epigenomics analysis. By performing an integrative meta-analysis of RNA-Seq and genome-wide DNA methylation data we will identify the key epigenetic enzymes driving the aberrant epigenome in specific cancer types. Further integration with ChIP-Seq data will help identify the regulatory networks which are disrupted in specific cancer types. In addition, a pan-cancer-wide approach will help elucidate universal patterns of epigenetic deregulation in cancer. Our ultimate goal is not only to improve our understanding of the cancer epigenome and identify key epigenetic drug targets, but to also provide a resource and open-source software for further exploration and hypothesis generation. This will help direct cancer epigenetic studies by identifying the most relevant epigenetic enzymes in cancer.

包括癌症在内的复杂疾病已成为现代社会的重要经济与社会负担，癌症研究的当务之急是加强对造成癌症发病的分子机制的了解。研究表明，许多癌症相关的驱动突变都位于表观遗传酶类相关基因上，然而对于表观遗传酶类异常如何影响癌症表观基因组从而造成调控网络破坏仍缺乏深入研究。.本项目将开发新型生物信息学方法，用以进行泛癌症范围内的系统表观基因组学研究：通过对RNA-Seq数据及全基因组DNA甲基化图谱进行整合分析，可识别造成特定癌症中表观基因组变异的关键酶类。对相关ChIP-Seq数据进一步进行整合分析将有助于识别在特定癌症类型中遭到破坏的调控网络。此外，利用泛癌症范围的分析手段将有助于识别在不同癌症类型中共有的表观遗传失调模式。其最终目的不仅在于加强我们对于癌症表观基因组学的理解，获得关键的表观遗传相关的药物靶点。本研究将有助于通过识别癌症关键的表观遗传酶类从而指导癌症表观遗传学研究。

我们的项目目标包括三个方面：（1）开发用于癌症系统表观基因组学的新型统计生物信息学方法，（2）将这些工具应用于表观基因组广泛关联研究，包括癌症表观基因组研究，并使用它们进行趋势分析，（3）帮助我们更好地理解癌症的系统级原理，包括识别癌症的因果表观遗传学改变和临床相关的生物标志物。这里总共有4个工作包（WPs）。在WP1中，我们旨在整理TCGA数据，并将FEM算法推广到较新的甲基化芯片。这已经成功实现，并且我们还扩展了FEM Bioconductor package。在WP2中，旨在使用DNA甲基化和RNA-Seq数据对癌症类型进行综合分析，以帮助鉴定在表观遗传水平上被破坏的癌症驱动基因。这项工作也已经成功完成。WP3旨在开发其他系统级分析方法，并且我们成功地实现了它，通过开发SEPIRA算法和Bioconductor 软件包。WP4是我们也已经成功完成的应用程序包。