Establishing A Prognostic Prediction Model Of Autophagy-Related Genes In Lung Adenocarcinoma Based On The TCGA Database

Objective:In the world,lung cancer is the most common cause of cancer death.As the most common type of lung cancer,lung adenocarcinoma accounts for about 40% of lung cancer patients,with an overall survival time of less than 5 years.Moreover,few biomarkers have been found to effectively predict the prognosis of patients with lung adenocarcinoma.Therefore,new therapeutic markers and targets are needed to achieve a better prognosis.Autophagy is considered to be an important catabolic process in eukaryotic cells,which causes lysosomes to degrade damaged,senescent or non-functional proteins and organelles.More and more evidences indicate that the interaction between autophagy and apoptosis is crucial in the pathophysiology of lung adenocarcinoma.The potential ability of autophagy to regulate cell death makes it a therapeutic target for cancer.This study will construct a prognostic prediction model of autophagy-related genes in lung adenocarcinoma through bioinformatics methods,and provide new ideas for the prognostic treatment of lung adenocarcinoma.Methods:1.Downloading the FPKM gene expression profile and clinical information of lung adenocarcinoma from the TCGA(The Cancer Genome Atlas)database.The "rtracklayer" package in the R language was used to isolate mRNA expression profiles and lncRNA expression profiles.2.Weighted gene co-expression analysis on the mRNA expression profile of lung adenocarcinoma was performed to screen out gene modules closely related to the occurrence of lung adenocarcinoma,and autophagy-related genes downloaded from the GSEA official website and gene modules were take the intersections to screen out common genes.3.The gene expression profile of the common gene was combined with the processed clinical information(survival status and survival time),and the cancer samples were randomly divided into a training set and a test set according to 50%,the training set was for constructing a prognosis predictive model,and the predictive model was combined with clinical data(age,gender,M,N,stage,T)for independent prognostic analysis.4.Analyzing the gene expression level of genes in the prediction model and seeking out related prognostic lncRNAs.Results:1.The mRNA expression profile of LUAD was analyzed by WGCNA,and it was found that pink module and brown module were closely related to the occurrence of lung adenocarcinoma.Among them,pink module is positively correlated with the occurrence of LUAD(the correlation is 0.71,P is 4e-91);the brown module is negatively correlated with the occurrence of LUAD(the correlation is-0.85,P is1e-164).2.In the training set,the single-and multi-factor COX regression analysis of the screened common genes were performed to obtain a prognostic prediction model,which included four gene signatures of TUBB6,TREM2,TRIM27 and ZC3H12 A.The results of the ROC curve and calibration chart of the training set and the testing set show that the prognostic prediction model has certain reliability and accuracy.The results of independent prognostic analysis show that the prediction model may become an independent prognostic factor of lung adenocarcinoma.3.After correlation analysis and single and multivariate COX regression analysis,five prognostic lncRNAs related to the genes in the model were obtained,namely AL121944.1,AL136304.1,FAM83 A.AS1,AC090559.1 and LINC00460.Conclusion:We have found four gene signatures(TUBB6,TREM2,TRIM27 and ZC3H12A)related to the prognosis of lung adenocarcinoma,and this prediction model may be an independent prognostic factor for lung adenocarcinoma.