基于生物信息学方法筛选并分析影响结直肠腺癌患者预后的铁死亡相关基因

Screening and analysis of ferroptosis-related genes impacting the prognosis of colorectal adenocarcinoma patients based on bioinformatics

目的利用生物信息学技术筛选并分析影响结直肠腺癌患者预后的铁死亡相关基因(FRG)。方法使用癌症基因组图谱(TCGA)数据库下载RNA测序数据, 共545例结直肠腺癌患者临床信息, 602例数据集。从FerrDb数据库下载FRG基因集。取FRG基因集与TCGA数据库基因集交集, 得到FRG表达数据集。使用R软件筛选结直肠腺癌组织与癌旁组织间差异表达FRG和预后相关基因, 获得影响结直肠腺癌预后的FRG。通过在线蛋白互作网络工具分析蛋白之间的互作关系及蛋白质表达的相关性。使用LASSO回归分析构建患者5年总生存率的预后风险模型, 计算TCGA数据库509例结直肠腺癌样本的风险值, 以中位风险值为临界值, 将患者分为高风险组(≥中位风险值)、低风险组(<中位风险值), 并绘制生存曲线。绘制依据FRG预后模型的风险值预测TCGA数据库中结直肠腺癌患者5年总生存率时间依赖的受试者工作特征(ROC)曲线, 使用Cox比例风险模型进行单因素和多因素生存分析, 筛选影响预后的因素。基于基因本体(GO)数据库和京都基因与基因组百科全书(KEGG)数据库对与结直肠腺癌预后相关的FRG进行通路富集分析。结果数据库中545例患者临床信息, 602个数据集中, 通过比较发现199个在结直肠腺癌中差异表达的FRG, 28个预后相关的FRG, 取交集后发现21个影响结直肠腺癌患者预后的FRG。DUOX2、NOX4、NOX1、DDIT3、JDP2、ATP6V1G2、ULK1、ATG3与WIPI1基因间可能有相关性;NOX4、NOX5、PLIN4和ATP6V1G2基因表达呈正相关, ULK1与MAPK1、MYB、FANCD2、ATG3、ATP5MC3基因表达呈负相关。使用LASSO回归分析, 最终筛选出15个FRG(ATP5MC3、NOX4、NOX5、ALOX12B、ATG3、WIPI1、MAPK1、MYB、AKR1C1、DDIT3、JDP2、ATP6V1G2、DRD4、SLC2A3、PLIN4), 构建风险模型, 风险值=NOX4×0.139-ATP5M3×0.108+NOX5×1.486+ALOX12B×0.475-ATG3×0.030-WIPI1×0.170-MAPK1×0.271-MYB×0.063+AKR1C1×0.021+DDIT3×0.186+JDP2×0.292+ATP6V1G2×0.777+ DRD4×0.294+SLC2A3×0.059+PLIN4×0.113。高风险组患者总生存较低风险组差(P<0.001), 低风险组5年总生存率为76.8%, 高风险组5年总生存率为48.2%。多因素生存分析显示年龄和风险值是预后的独立影响因素。使用预后相关FRG构建的风险模型能够较好地预测患者5年总生存率(曲线下面积0.728)。高、低风险组患者中的差异表达基因可能与细胞外基质组成和细胞外结构构成、局部黏附等基因通路有关。结论依据筛选的FRG构建的预后风险模型可较好评估结直肠腺癌患者预后, 这些FRG有望成为结直肠腺癌预后相关新的候选生物标志物。

Objective To screen and analyze ferroptosis-related genes(FRG)impacting the prognosis of colorectal adenocarcinoma patients based on bioinformatics.Methods RNA sequencing data including the clinical information of 545 colorectal adenocarcinoma patients and 602 data sets were downloaded from the Cancer Genome Atlas(TCGA)database.FRG gene sets were downloaded from FerrDb database.FRG expression dataset could be obtained after taking the intersection between FRG gene sets and TCGA database gene sets.Differentially expressed FRG and prognosis-related genes between colorectal adenocarcinoma tissues and the adjacent tissues were screened by using R software,and finally FRG influencing the prognosis of colorectal adenocarcinoma were obtained.According to protein-protein interaction networks,the interaction and the expression association of proteins were analyzed.LASSO regression analysis was used to build a risk model for patients'5-year overall survival rate.The risk value was calculated for 509 colorectal adenocarcinoma samples in the TCGA database,and then the median risk value was taken as the cut-off value.All patients were divided into the high-risk group(≥median risk value)and the low-risk group(<median risk value),and the survival curves of the two groups were drawn.The receiver operating characteristic(ROC)curve was drawn for predicting the 5-year overall survival rate of colorectal adenocarcinoma patients in a time-dependent way in TCGA database according to the risk value of FRG prognosis model.Cox proportional risk model was used to make univariate and multivariate survival analysis in order to screen factors affecting the prognosis.The pathway enrichment analysis of prognosis-related FRG of colorectal adenocarcinoma was performed based on gene ontology(GO)database and Kyoto Encyclopedia of Genes and Genomes(KEGG)database.Results The clinical information of 545 patients and 602 datasets were extracted from the database.A total of differential expressed 199 FRG in colorectal adenocarcinoma and 28 prognosis-related FRG were identified.After taking the intersection,21 FRG affecting the prognosis of colorectal adenocarcinoma patients were identified.DUOX2,NOX4,NOX1,DDIT3,JDP2,ATP6V1G2,ULK1,ATG3 were probably associated with WIPI1;expressions of NOX4,NOX5,PLIN4 were positively correlated with ATP6V1G2,while the expression of ULK1 was negatively correlated with MAPK1,MYB,FANCD2,ATG3 and ATP5MC3.LASSO regression analysis showed that 15 FRG were finally screened out(ATP5MC3,NOX4,NOX5,ALOX12B,ATG3,WIPI1,MAPK1,MYB,AKR1C1,DDIT3,JDP2,ATP6V1G2,DRD4,SLC2A3,PLIN4),and the risk model was constructed by calculating the risk value,and the risk value=NOX4×0.139-ATP5M3×0.108+NOX5×1.486+ALOX12B×0.475-ATG3×0.030-WIPI1×0.170-MAPK1×0.271-MYB×0.063+AKR1C1×0.021+DDIT3×0.186+JDP2×0.292+ATP6V1G2×0.777+DRD4×0.294+SLC2A3×0.059+PLIN4×0.113.The overall survival of patients in the high-risk group was worse than that in the low-risk group(P<0.001).The 5-year overall survival rate was 48.2%in the high-risk group and 76.8%in the low-risk group.Multivariate survival showed that the age and risk value were independent affecting factors of the prognosis.ROC curves revealed that the risk model constructed by using prognosis-related FRG could well predict the 5-year overall survival rate of patients(the area under the curve was 0.728).The differential expressed genes of both groups may be associated with genetic pathways such as extracellular matrix composition,extracellular structure composition and focal adhesion.Conclusions The prognostic risk model constructed by the screened FRG can better evaluate the prognosis of colorectal adenocarcinoma patients.These FRG are expected to become new candidate biomarkers related to the prognosis of colorectal adenocarcinoma.