基于高通量表观基因组图谱的癌症分子分型的系统分析

Accurate molecular subtypes in tumor can provide effective guidance for tumor diagnosis and treatment. High throughput epigenomic data facilitate our systematic understanding of cancer molecular typing based on epigenetic modifications and the subtype-specific epigenetic regulation mechanisms. This project is carried out in breast cancer model according to the following research process: cancer information database→ cancer molecular typing by epigenetic modification→ analysis of subtype-specific epigenetic regulation mechanism→ identification of subtype specific epigenetic marker. We collect DNA methylation, histone modifications, lncRNA and gene expression profiles in various breast cancer samples detected by multiple experiment technologies, and build a comprehensive breast cancer database. Next, limited clustering algorithm is developed and used to classify breast cancer subtypes based on epigenetic modifications. By pattern recognition algorithms, we systematically analyze the molecular mechanism by which aberrant epigenetic modifications mediate different subtypes of cancer. Then, we carry out comparative analysis of subtype-specific epigenetic state profiles, and identify biomarkers for molecular diagnosis and targeted treatment of breast cancer. At last, integrated analysis platform is built and applied to molecular typing of other cancers. The mining of cancer subtype-specific epigenetic features and regulation mechanism based on our technology advantage in bioinformatics and the combination of bioinformatic and experimental methods should provide new strategies for molecular typing, targeted therapy, and molecular mechanism of cancer.

肿瘤精确的分子分型可为肿瘤诊断和治疗提供有效的指导，高通量实验产生的表观基因组数据促进系统地基于表观遗传修饰的癌症分子分型及亚型特异致病机制的阐明。本课题以乳腺癌为模型，依照癌症信息数据库→表观遗传修饰的癌症分子分型→亚型特异的表观遗传调控机制分析→亚型特异的表观遗传标记物识别的研究思路，收集多种实验技术测定乳腺癌多样本的DNA甲基化、组蛋白修饰、lncRNA和基因表达谱等信息，构建综合的乳腺癌信息库，开发有限聚类算法基于表观遗传修饰分类乳腺癌亚型；应用模式识别算法系统挖掘表观遗传修饰异常介导不同亚型的分子模式，比较分析亚型表观遗传状态图谱，发现乳腺癌亚型分子诊断和靶向治疗的标志物；构建分析技术平台并将该平台推广至其他癌症的分子分型。利用课题组的生物信息学技术优势，结合生物信息学与实验研究手段，挖掘癌症亚型特异的表观基因组特征、调控机制，可提供癌症的分子分型、靶向治疗和分子机制研究新策略。

肿瘤精确的分子分型可为肿瘤诊断和治疗提供有效的指导，高通量实验产生的表观基因组数据促进系统地基于表观遗传修饰的癌症分子分型及亚型特异致病机制的阐明。本课题针对癌症亚型相关的表观遗传修饰分子分型特征及表观遗传调控因子介导分子亚型的作用模式的挖掘，构建了系统的癌症表观遗传状态图谱，从全基因组尺度分析癌症亚型产生中DNA甲基化与其他多种表观遗传调控元件的作用机制及生物学功能。重点利用高通量的表观基因组数据资源开发数据挖掘方法、软件平台（包括DevMouse，MetaImprint，SEA，SMART和DiseaseMeth version 2.0等），这些方法和软件提供了从高通量数据中挖掘癌症亚型间的表观遗传变化提供了工具。项目通过分析细胞系与组织间的DNA甲基化特征，对乳腺癌进行了精细的分子分型，发现不同分型的预后存在着差别; 利用典型相关和热点分析的算法挖掘到乳腺癌预后相关的DNA甲基化标记。基于分子生物学网络分析的方法识别了不同癌症的DNA甲基化标记。本项目的研究揭示了表观遗传改变在肿瘤内和肿瘤间异质性产生中的重要作用，研究成果将为癌症的分子分型、分子诊断和靶向治疗等众多方面提供新的思路及实验验证的基础。