胃癌转移相关的基因标志的可重复性评价及再识别

Tumor metastasis status provides important evidence for customizing patients’ treatment. However, the metastasis-related gene signatures for cancers including gastric cancer have low reproducibility. The reasons include the following: (1) Imaging-based and pathologic examination in detecting tumor metastasis produces high rates of false positive and negative. (2) The high variation of gene measurements and the uncertainty of the quality of clinical samples. (3) The proportions of tumor epithelial cell within different tumor sites are very different for the same tumor tissue. Thus, in this study, we plan to make the following analyses: (1) Based on the within-sample relative expression orderings (REOs) of genes, we identify the predictive signature for micro-metastasis and postoperative recurrence risk for early gastric cancer. This REOs-based signature is insensitive to molecular measurement variations, experimental batch effects and partial RNA degradation. (2) Using this predictive signature, we exclude the potential false negative metastasis samples which are predicted at high recurrence risk from the samples without lymphatic and distant metastasis and the potential false positive metastasis samples which are predicted at low recurrence risk from the samples with lymphatic and/or distant metastasis, respectively. Then, we identify the metastasis-related signature genes with transcriptomic, epigenomic and genomic alteration. (3) To collect and sequence the samples from different tumor sites for the same tumor patient, we evaluate and optimize the mutation calling algorithm which is used for evaluating the proportion of tumor epithelial cell based on the whole-exome sequencing data and adjusting the influence of this proportion on mutation calling. (4) Re-identify the reproducibly metastasis-related multi-omics signatures in the eight main types of cancer like gastric cancer. (5) Construct the model of metastatic feature pathways regulated by the genes with somatic mutation or copy number alteration in the genome or abnormal DNA methylation.

肿瘤转移状态是制定肿瘤治疗方案的重要依据，但肿瘤转移相关的基因标志的可重复性很低，原因包括：（1）肿瘤转移的影像、病理诊断的假阳性和假阴性率高；（2）基因检测高变异与临床样本质量不确定；（3）肿瘤组织取样部位的癌细胞占比变化。因此，本课题拟（1）基于样本内基因表达水平的相对高低秩序关系，识别早期胃癌的微转移与术后复发风险预测标志，该类标志对分子检测的批次效应及RNA部分降解等不敏感；（2）利用该标志排除可能的转移假阴性（复发高风险的N0M0样本）与假阳性（复发低风险的N+或M1样本）的样本，再识别转移相关的转录组、表观组与基因组特征基因；（3）对一组肿瘤组织多部位取样并测序，优化根据外显子测序数据估计癌细胞占比及矫正其影响的突变识别算法；（4）重新识别胃癌等8种主要癌型的转移相关的可重复的各类组学标志物；（5）建立基因组突变/CNA及甲基化异常功能模块调控转移特征通路的模型。

剖析转移的分子机制一直是改善癌症患者预后的热点研究方向。然而，转移的临床诊断具有较高的假阴性，导致转移和非转移样本间差异的生物学信号较弱。本项目提出借助转录组预后标志对转移进行重分类能够提高差异信号的假设，以胃癌为主要研究对象，识别预测转移风险标志，据此挖掘可重复的与转移相关的表观组和基因组等特征。.首先，针对定量转录组标志受批次效应影响的技术问题，本项目提出基于基因表达秩序关系开发定性标志的方法，并应用于II-III 期胃癌样本。本项目识别了一个包含19个基因对的定性转录预后标志，其预测II-III 期胃癌患者转移风险的能力在两套数据集中得到了验证。借助标志进行转移分类，我们识别了可重复的与转移相关的表观组和基因组等特征，为揭示胃癌转移的分子机制提供重要帮助。.接下来，将本项目开发的方法应用于肺腺癌研究中时发现，在RNA测序数据中，不同基因注释处理会导致转录组标志对患者的风险预测发生变化。针对该问题，我们选择在不同基因注释版本里稳定的基因，开发一个包含40个基因对的肺腺癌定性预后标志。标志的稳定性在临床收集的I期石蜡包埋样本中得到了验证。我们利用开发的转录标志对肺腺癌的转移风险进行分类，识别两组间差异的克隆和亚克隆突变，开发了一个稳健的预测预后以及免疫疗效的突变标志。.最后，针对胃癌高度异质性的问题，本项目将依赖于基因表达秩序关系开发的个体化差异分析方法应用于1090例胃癌样本中，识别了25个在90%以上样本差异表达的基因，其失调频率在448例配对胃癌样本中得到了验证。我们收集了24例配对胃癌和癌旁样本进行RNA测序，验证了它们普遍差异的情况。蛋白互作网络分析支持这些蛋白参与了癌相关功能。.综上所述，该项目开发的定性方法成功应用于胃癌、结肠癌、肺癌等多个癌型，为癌症发生和转移的分子机制研究提供新的思路；识别的预测标志及功能基因为癌症的诊断和治疗提供分子基础。本项目已发表SCI论文24篇。

内容获取失败，请点击重试