基于脂代谢相关基因预测局部晚期直肠癌新辅助放化疗疗效

Predicting the efficacy of neoadjuvant chemoradiotherapy for locally advanced rectal cancer based on lipid metabolism-related genes

目的探讨基于脂代谢相关基因(LMRG)预测局部晚期直肠癌(LARC)新辅助放化疗疗效的价值。方法于基因表达数据库获得接受新辅助放化疗的LARC的全基因组表达数据GSE46862,进行差异表达分析以获得差异表达基因。于分子标签数据库(MSigDB数据库)搜集LMRG并与差异表达基因取交集获得差异表达的LMRG。基于最小绝对收缩和选择算子(LASSO)回归、支持向量机递归特征消除(SVM-RFE)、随机森林(RF)三种机器学习算法筛选获得候选LMRG。采用基因本体论(GO)与京都基因与基因组百科全书(KEGG)分析进行功能富集分析以获得潜在的功能与作用通路。采用受试者操作特征(ROC)曲线分析评估候选LMRG预测LARC新辅助放化疗疗效的准确性。结果共筛选出8个候选LMRG(ALOX5AP、FADS2、GALC、PLA2G12A、AGPAT1、AACS、DGKG、ACSBG2),这些LMRG主要涉及脂质代谢相关生物进程,并参与调控多个重要的脂质代谢相关信号通路。此外,这8个候选LMRG拥有较高的预测LARC新辅助放化疗疗效的曲线下面积(AUC)值。结论基于3种机器学习算法鉴定出的8个LMRG拥有较高的预测LARC新辅助放化疗疗效的准确性,可为寻找LARC术前新辅助放化疗疗效评估的分子标志物及潜在的治疗靶点提供线索。

Objective To investigate the value of lipid metabolism-related genes(LMRG)for predicting the efficacy of neoadjuvant chemoradiotherapy in locally advanced rectal cancer(LARC).Methods GSE46862,a genome-wide expression data of LARC treated with neoadjuvant radiotherapy,was obtained from the Gene Expression Database,and differential expression analysis was performed to obtain differentially expressed genes.The LMRG were collected from the MSigDB database and intersected with differentially expressed genes to obtain differentially expressed LMRG.Candidate LMRG were identified based on three machine learning algorithms including least absolute shrinkage and selection operator(LASSO),support vector machine-recursive feature elimination(SVM-RFE),and random forest(RF).Functional enrichment analysis was performed using gene ontology(GO)and Kyoto Encyclopedia of Genes and Genomes(KEGG)analysis to obtain potential function and involved pathways.The accuracy of the candidate LMRG in predicting the efficacy of neoadjuvant chemoradiotherapy for LARC was assessed using receiver operating characteristic(ROC)curve analysis.Results A total of eight candidate LMRG(ALOX5AP,FADS2,GALC,PLA2G12A,AGPAT1,AACS,DGKG,ACSBG2)were screened which were mainly involved in biological processes related to lipid metabolism and were involved in the regulation of several important lipid metabolism-related signaling pathways.In addition,these eight candidate LMRG possessed high area under the ROC curve(AUC)for predicting the efficacy of neoadjuvant chemoradiotherapy for LARC.Conclusion The eight LMRG identified based on three machine learning algorithms had high accuracy in predicting the efficacy of neoadjuvant chemoradiotherapy for LARC,providing clues to identify molecular markers and potential therapeutic targets for preoperative neoadjuvant radiotherapy evaluation of LARC.