拷贝数变异（cnv）广泛存在基因组中并通过剂量效应影响基因表达，进而导致癌症的发生和发展。整合cnv与基因表达的研究，有助于从癌症众多的分子变化中识别出处于驱动地位的driver基因。然而，一些基因在正常组织和癌组织都发生cnv，但表型效应却截然不同。目前导致此现象的潜在机制尚不明确。我们的预实验发现，正常人群中的cnv区域比癌相关的和非cnv区域显著更多地受到mirna和tf的靶向。据此，提出"人体内存在抑癌的平衡子mirna与tf协同纠错系统"的假设，推测mirna和tf共调控通路的失调是引发癌症的原因之一。本项目拟整合cnv、基因表达及转录调控等多组学信息，构建三维调控网络，开发研究抑癌机制的新框架，识别乳腺癌的抑癌平衡子mirna与tf及其协同调控通路，并推广研究其他癌症，建立综合分析平台，阐明调控cnv基因转录过程的机制及其对癌症的影响，为有效防治癌症及抗癌药物的研发提供新线索。

copy number variations (cnvs) which are pervasive in human genome, influence gene expression through dosage effects, which would cause the occurrence and development of cancers. integrating cnvs and gene expression will go far towards distinguishing the driver genes in cancer from numerous aberrant molecules. however, some genes show cnvs in both normal and tumor tissues but with distinct phenotypic effects. the underlying mechanism of this phenomenon is unclear by far. by preliminary tests, we found that the cnv regions in normal people were more frequently targeted by mirnas and tfs than cancer-related or non-cnv regions significantly. accordingly, we proposed a hypothesis "a tumor-suppressing synergetic correction system of mirna and tf exists in human bodies" and speculated that the dysregulation of mirna-tf co-regulatory pathways is one cause of cancer. this project will devote to construct a three-dimensional regulatory network by integrating multiple-omics information of cnvs, gene expression and transcriptional regulation, develop a novel framework for analyzing anticancer mechanism, identify balancer mirna/tf and their synergetic regulatory pathways in breast cancer, and develop a comprehensive analytical platform for other cancers analysis. our work aims to explain the mechanism of cnv gene transcription process and its effects to cancer, which will provide valuable clues for quick and efficient cancer prevention and cure as well as the development of anticancer drugs.

本项目通过整合分析拷贝数变异、基因表达、非编码rna及转录调控等多组学信息，开发出筛选癌症剂量抵抗与剂量敏感基因的计算框架。基于tf和mirna介导的ffl，建立了mirna与tf协同调控剂量效应基因的三维调控网络。通过对网络特性研究，分析了ffl对剂量抵抗基因和敏感基因的调控作用，识别出关键调控子以及癌症预后相关网络motif。同时，联合tf主导的转录调控以及mirna主导的转录后调控所构成的协同调控模序，研究了癌症状态下pcg与lncrna在拷贝数、表达和剂量效应方面的异同，探讨了各种协同调控的模序对pcg和lncrna剂量效应的影响。研究发现，剂量敏感基因与细胞周期、dna代谢等癌症基本过程相关，而剂量抵抗基因与癌症免疫反应、细胞死亡与凋亡等过程相关。剂量敏感度高的基因拥有更多的mirna结合位点，受到更多mirna的靶向调控。mirna表达的改变可以影响原癌基因与肿瘤抑制基因的剂量敏感性，可作为癌症的诱因或平衡子。本项目研究发现，mir-100-5p和let-7b-5p高表达样本有更长的生存时间而mir-375和mir-429低表达的样本具有更长的生存时间，可以作为乳腺癌的预后靶标。本项目还识别出了卵巢癌新的候选癌基因如pdpn、epha2等及一些关键调控子mir-16-5p, mir-98-5p, myb 和 hoxa5。并识别出能够把卵巢癌样本显著分成高、低风险组可以作为卵巢癌的预后标记的癌相关ffls：mir-130a-3p-stat1-cdkn1a与mir-15b-5p-myb-igf1r。另外，通过构建肺腺癌cerna网络，我们还识别出了可以作为肺腺癌预后标记的cernas：ensg00000240990- hoxa10-hsa-let-7a/b/f/g-5p此外，本项目利用tcga与ccle中对应的基因表达谱、拷贝数变异谱和突变数据，计算样本间相关性；在功能层面结合样本间go term网络相似性，为8种癌症筛选出最适合作为体外模型的细胞系。