肿瘤相关子通路的多维组学个体化分析:方法与应用

For mRNA, miRNA and DNA methylation profiles of tumor, the common analyses were focused on identifying inter-group (such as cancer and normal groups) differential genes and these genes involved functional pathways. However, such differential genes may be the mixture signals of differential genes from different subtypes compared with their normal control (mosaic phenomenon) and cannot reflect the real differential genes occurred in individuals. We firstly showed that the within-sample relative expression orderings of genes are highly stable in a particular type of normal tissues but are widely disrupted in the corresponding cancer tissues (the DNA methylation level and miRNA expression level existed the same phenomenon). Based on this phenomenon, we developed algorithms for individual-level analysis of differential expression of mRNAs (genes that differentially expressed in one cancer sample compared with its precancerous normal tissue), DNA methylation and miRNA. Based on individual-level analysis of differential genes, we will develop individual-level subpathway analysis methods of mRNA, DNA methylation and miRNA (target genes). We will also develop an algorithm to identify common and subtype specific mutation subpathway for a particular tumor. Using the multi-omics data of lung cancer from TCGA, we will identify the common and subtype specific multi-omics subpathways for lung cancer, and build the upstream genomic and epigenomic damage models of the common and subtype specific transcriptional subpathways.

对肿瘤mRNA、miRNA和DNA甲基化谱，最基础的分析是识别癌与正常组织间的差异基因及其富集的功能通路。然而，两组间存在统计差异的基因与通路可能是不同亚型的差异基因的混合信号（“马赛克”现象）,而非反映真实地发生于任何肿瘤个体的异常。根据基因表达相对秩序关系在一特定正常组织中高度稳定而在相应癌组织中广泛逆转这一生物学现象（DNA甲基化和miRNA也存在类似的现象），我们已开发了识别个体化差异mRNA（即每个癌组织样本相对其癌变前的正常状态差异的基因）、DNA甲基化和miRNA的算法。本课题拟以个体化差异基因分析为基础，发展mRNA、DNA甲基化和miRNA（靶基因）的个体化子通路分析算法，并发展识别特定癌型共性和亚型特异的突变子通路的分析算法。利用TCGA肺癌的多维组学数据，识别肺癌共性和亚型特异的多维组子通路，并建立调控肺癌共性及亚型特异的转录组子通路的上游基因组及表观组损伤调控模型。

对肿瘤mRNA谱，最基础的分析是识别癌与正常组织间的差异基因及其富集的功能通路。然而，两组间存在统计差异的基因与通路可能是不同亚型的差异基因的混合信号（“马赛克”现象）,而非反映真实地发生于任何肿瘤个体的异常。. 为了解决该问题，针对mRNA分析本课题设计了两种方法来识别特定癌型共性的子通路：（1）通过改进个体化差异mRNA分析算法识别在显著多的疾病样本中上调和下调的差异基因，将这些基因富集的通路定义为疾病共性的通路，且该方法可以应用于仅有疾病样本（单表型）的基因表达谱来识别疾病相对正常对照的差异基因；（2）基于个体化差异表达基因分析方法RankComp和配对样本数据来识别在90%的胃癌样本中差异表达的基因，并以这部分基因作为种子识别胃癌共性的子通路，追溯其上游基因组及表观组损伤。另外，根据胃癌病人基于氟尿嘧啶放化疗的预后信息构建了识别不同预后亚型的标志，并识别了亚型特异的子通路。. 针对突变子通路分析，本课题基于肿瘤样本的突变谱开发了识别特定癌型共性和亚型特异突变子通路分析算法PathMG，并将该算法分别应用于肺癌和结直肠癌中识别共性和亚型特异的突变子通路。. 综上所述，该项目基于肿瘤样本的突变谱和mRNA表达谱已经开发了特定癌型共性和亚型特异的子通路分析算法，可为癌症发生机制的研究提供新的思路，为癌症诊断和分型提供潜在的标志。已发表SCI论文7篇。