肾细胞癌亚型的特征基因筛选及识别研究  

Screening and Identification of Characteristic Genes of Renal Cell Carcinoma Subtypes

作　　者：付继鹏 韩振波 李晓琴[1] 王猛[1] 

机构地区：[1]北京工业大学环境与生命学部,北京

出　　处：《生物物理学》2023年第1期1-16,共16页Biophysics

摘　　要：目的:肾细胞癌(Renal Cell Carcinoma, RCC)不是单一疾病,而是几种具有不同遗传驱动因素、临床病程和治疗反应的组织学定义的癌症。为了研究不同肾细胞癌亚型的分子特征,本文提出了一个筛选肾细胞癌亚型分类的流程图,并对亚型的特征基因进行了分析。方法:基于DNA甲基化和基因表达数据,综合统计学方法和机器学习方法,筛选了与肾细胞癌三种亚型相关的特征基因集,并构建了一个肾细胞癌亚型分类模型。结果:本文筛选得到了6个分类特征基因;构建的分类器准确性达到96.6%,精确性和敏感性分别是93.4%、94.7%;独立检验集的准确性为93.1%。对肾细胞癌亚型相关的特征基因集进行富集分析发现:肾透明细胞癌主要与免疫系统的负调节和白细胞增殖与黏附等相关通路有关,肾乳头状细胞癌和肾嫌色性细胞癌都与肾脏和肾脏系统的发育有关。结论:本文通过构建分类模型实现了肾细胞癌亚型的三分类,结果对了解肾细胞癌的亚型形成机制及分类诊断治疗具有一定指导意义。Purpose: Renal cell carcinoma (RCC) is not a single disease, but several histologically defined can-cers with distinct genetic drivers, clinical course, and response to therapy. To investigate the mo-lecular characteristics of the different RCC subtypes, a flowchart for the classification of the subtypes of RCC is presented, and the signature genes of the subtypes are analyzed. Methods: Based on DNA methylation and gene expression data, integrated statistical methods and machine learning meth-ods, signature gene sets associated with the three subtypes of renal cell carcinoma were screened, and a model for renal cell carcinoma subtype classification was constructed. Results: In this paper, six classification signature genes were obtained;The accuracy of the constructed classifier reached 96.6%, and the precision and sensitivity were 93.4%, 94.7%;The accuracy of the independent test set was 93.1%. By enrichment analysis of the characteristic gene set related to renal cell carcinoma subtypes, it was found that kidney renal clear cell carcinoma is mainly associated with negative regulation of the immune system and related pathways, such as leukocyte proliferation and adhe-sion. Both Kidney renal papillary cell carcinoma and Kidney Chromophobe are associated with the development of the kidney and the renal system. Conclusion: In this paper, we achieved the three classification of renal cell carcinoma subtypes by constructing a classification model, and the results will be helpful for understanding the mechanism of renal cell carcinoma subtype formation and for classification, diagnosis, and treatment.

关 键 词：肾细胞癌 肾透明细胞癌 临床病程 DNA甲基化 特征基因 分类模型 基因表达数据 分子特征 

分 类 号：R73[医药卫生—肿瘤]