整合生物代谢网络和新一代测序数据识别卵巢癌相关基因及风险通路的研究

Identifying ovarian cancer (OvCa) genes and risk pathways could help uncover disease pathogenesis of susceptible genes and metabolic dysfunctions during disease process, which could further aid in drug target identification and effective diagnosis/treatment decision support. Previously, we observed on the copy number variation data in case and control samples of about 20 cancer types, and found ovarian cancer is a high risk cancer that driven by copy number variations. And we also noticed that due to less overlaps of tested samples and small sample sizes, there is a large discrepancy of predicted cancer genes from different genomic data. It is extremely essential of getting cross-validated information between multiple high-throughput genomic data on multiple platforms when we work on accurate identification of cancer susceptible genes and risk pathways. In addition, traditional risk pathway identification methods often elaborate in search of single risk pathway related to OvCa and cannot illustrate the "cross-talk" regulatory mechanisms between multiple biological risk pathways, thus leading to the failure in better understanding of the global metabolic dysfunctions of biological processes in cancer patients. In this project, we will focus on ovarian cancer and analyze the next generation sequencing omics data provided by the Cancer Genome Atlas (TCGA) project. In order to identify susceptible genes and risk pathways, we will start from copy number variation data and then integrate it with DNA methylation and gene expression data to find cross-validated information between multiple omics datasets using both bioinformatics and computational systems biology techniques. This study will help to deepen our understanding of the underlying mechanisms of global metabolic dysfunctions during the process of ovarian cancer onset and progression.

识别卵巢癌(OvCa)基因和风险通路，可以揭示疾病基因致病机制和疾病发生过程中的代谢紊乱机制，帮助发现药物靶点和有效确立诊断治疗方案。我们前期分析了20多种癌症样本的拷贝数数据，发现并证实了卵巢癌属于拷贝数变异驱动的高危癌症。我们还注意到不同的基因组数据预测基因集合重复性差，这与不同基因组数据的样本不一致及整体样本数量太小有关。对高通量的癌症样本数据进行分析并获取多平台相互印证的数据信息，对于癌症致病基因和风险通路的准确确定至关重要。另外，当前识别OvCa风险通路的研究一般只能发现单一风险通路，不能获得多个生物风险通路之间的协同调控机制。我们将利用"癌症基因组蓝图"计划提供的新一代测序数据，从拷贝数数据出发，采用生物信息和计算系统生物学分析技术，整合DNA甲基化和表达数据，识别OvCa致病基因，并鉴别与之密切相关的风险生物代谢通路。本项目实施将为卵巢癌综合代谢紊乱机制的探索提供一种新途径。

癌症等复杂疾病的发生伴随着多个分子层面的生物途径的改变，其中生物代谢网络是研究癌症等复杂疾病代谢紊乱机制最直接的背景，是研究癌症最为有效的网络资源之一。当前，利用生物信息学方法，挖掘疾病相关的基因、酶及药靶、研究表型相似性、识别与疾病相关模块和风险通路等，已成为研究复杂疾病机制的热点。但现有的生物代谢网络数据存在数据完整性不强、数据重叠率不高和存在一些错误描述的互作关系等诸多不足。因此，有必要整合现有数据资源，以形成更为完善的人类代谢网络，实现更全面揭示癌症等复杂疾病发生发展机制的目的。本课题组整合分析现有的人类代谢通路数据资源和蛋白质（基因）功能、定位相关信息，自主研究设计了多种重构策略模型，获得了功能注释较为完善且具有较高置信度的整合人类代谢网络数据集，并利用PHP+MySQL等技术，初步实现了具有多关键词关联性分析、疾病相关模块识别和网络可视化分析等特色的代谢物综合数据分析平台--人类代谢和通路功能仓库(HMFR) 的构建。与此同时利用多平台新一代测序技术获得的高通量基因组遗传变异数据以及其他多组学的复杂疾病（包括卵巢癌OvCa）数据资源，构建具有针对性的生物网络背景，创新性地提出相应的疾病基因、致病通路（疾病风险模块）、潜在药物靶点的高效识别算法，实现了OvCa致病基因及相关风险通路的可视化查询及分析功能。通过对于癌风险lncRNAs和其显著调控的癌风险通路的识别研究、基于生物网络优选复杂疾病致病基因方法和复杂疾病潜在药物靶点识别方法拓展性地研究，有望为癌症等复杂疾病致病机制的探索提供新的有效的途径，对研究挖掘特定表型复杂疾病相关的分子靶点具有重要指导意义