单细胞测序联合孟德尔随机化分析鉴定前列腺癌预后相关基因及分析

Identification of prognostic genes in prostate cancer by single-cell sequencing combined with Mendelian randomization

目的:结合患者临床信息,通过单细胞测序联合多组学分析,识别前列腺癌发生发展中的关键基因,并利用转录组学进一步验证,鉴定出前列腺癌预后相关基因。方法:前列腺癌患者的单细胞测序数据来自GSE156632数据库中的4例患者,利用R语言进行细胞聚类和注释,随后挑选出上皮细胞进行差异分析,使用Hiplot网站对鉴定出的差异基因进行富集分析,随后利用UKB(UK Biobank)数据库下载基因对应的表达数量性状基因座(expression quantitative trait locus,eQTL)的单核苷酸多态性(single nucleotide polymorphism,SNP)位点,同时从TCGA(The Cancer Genome Atlas)和GEO(Gene Expression Omnibus)数据库下载了前列腺癌患者临床数据以及对应的基因表达数据。采用单因素COX回归分析观察孟德尔随机化后基因对患者预后的影响。结果:利用Seurat包对4例患者的单细胞数据进行质控和整合,在进行质控和细胞类型鉴定后,挑选出了上皮细胞亚群进行差异分析,共得到1566个基因,随后进行了富集分析,结果显示差异基因可能富集于Ras信号通路,细胞凋亡与氧化磷酸化等信号通路,随后使用1566基因进行了孟德尔随机化,得到了74个可能的因果基因,使用TCGA-PRAD和GSE116918对这74个基因进行了单因素COX回归分析,发现FAM3B,JUNB,TMEM59,KRT5这4个可能的相关基因,最后比较了孟德尔随机化和单因素COX回归的结果,发现KRT5可能是对前列腺癌最有影响的基因。结论:这些结果表明,FAM3B,JUNB,TMEM59,KRT5可能在前列腺癌进展中起到作用,KRT5可能成为前列腺癌的预后预测因子和治疗靶点。

Objective:To identify the key genes involved in the development and progression of prostate cancer(PCa)and those associated with the prognosis of the malignancy.Methods:We obtained the single-cell sequencing data on 4 cases of PCa from the GSE156632 database.Using R language and the Seurat package,we performed cell clustering and annotation,selected the subpopulations of epithelial cells for differential analysis after quality control and cell type identification,and conducted enrichment analysis of the identified differential genes using the Hiplot website.Then we downloaded the single nucleotide polymorphism(SNP)loci corresponding to the expression quantitative trait loci(eQTL)of these genes from the UK Biobank(UKB)database,and the clinical data and corresponding gene expression data on PCa patients from The Cancer Genome Atlas(TCGA)and Gene Expression Omnibus(GEO),followed by univariate COX regression analysis of the impact of the genes on the prognosis of the patients after Mendelian randomization.Results:A total of 1566 genes were identified and subjected to enrichment analysis,which indicated that the differential genes might be enriched in the Ras,apoptosis and oxidative phosphorylation signaling pathways.Subsequent Mendelian randomization revealed 74 potential causal genes among the 1566 genes,and univariate COX regression analysis of the 74 genes identified 4 possibly related genes FAM3B,JUNB,TMEM59,and KRT5.Comparison of the results of Mendelian randomization and univariate COX regression showed that KRT5 might be the most important gene influencing PCa.Conclusion:FAM3B,JUNB,TMEM59 and KRT5 may play a role in the progression of PCa,and KRT5 may potentially serve as a prognostic predictor and therapeutic target for the malignancy.