Construction Of A Prognostic Risk Model For Esophageal Cancer Based On Variable Shear Events

Esophageal cancer is the ninth most common cancer worldwide and the sixth leading cause of cancer death.It can be divided into two histological subtypes: esophageal squamous cell carcinoma(ESCC)and esophageal adenocarcinoma(EAC).Due to its strong concealment,it is not easy to be detected,so the early symptoms are not obvious and it is difficult to be diagnosed in advance.Therefore,most patients have been diagnosed in the middle and late stages,and the therapeutic prognosis is poor at this time.Therefore,it is urgent to select efficient detection methods to improve the early diagnosis rate of esophageal cancer,so as to enable early detection and treatment of patients and reduce the risk of death.On the one hand,a large number of studies have shown that variable shear events are closely related to the occurrence and development of esophageal cancer: in tumor cells,variable shear events can affect the proliferation,metastasis,apoptosis and other important biological processes of tumor cells by affecting multiple genes and signaling pathways.On the other hand,as far as the current medical level is concerned,no mature and accurate early detection method of esophageal cancer has been developed.However,with the hot research on tumor-related biomarkers and the rapid development of the new generation of information technology in recent years,more and more scholars try to apply bioinformatics methods to tumor researches.The hope is to find biomarkers related to tumor through this interdisciplinary fusion.In this study,gene expression data,clinical data and variable shear data which used as original data were obtained from open databases,and variable screening and Cox regression model construction were combined with statistics and machine learning methods.Variable shear events related to esophageal cancer survival were found as scheduled,so as to provide new targets for the detection and treatment of esophageal cancer,and indicates the direction of developing corresponding targeted drugs.In this study,effective clinical data and gene expression data of 185 patients were obtained from the TCGA database.Combined with variable Splice data from the TCGA Splice database,the prognostic risk model of esophageal cancer was constructed using R software.Finally,regulatory networks of variable splice and splice factors were constructed using Cytoscape software.This study firstly collates the data,and then puts the collated data into the single-factor Cox regression model to initially explore the variable shear events related to survival.Then these events are put in the random forest model,Lasso regression model and elastic network model for variable screening,and then construct the multi-factor Cox regression model respectively.Finally,the survival related variable shear events screened by the three methods were obtained.The results showed that the types and numbers of variable shear events screened by Lasso and elastic network were identical,with 10 selected by both,while only 6variable shear evens were selected by random forest.After discussion and analysis,the variable shear events that screened by Lasso regression and elastic network are selected as the final biomarkers based on the principle of minimizing patient loss.So the final 10 biomarkers in our study were EGLN3|27150|AD,C19orf82|47381|ES,KALRN|66523|AT,TRIM16L|39638|ES,UBC|25166|RI,CLASRP|50387|AP,COX6C|84682|AD,SVIL|11108|A P,CCNB2|30929|ES and POMZP3|80186|ES.On this basis,we conducted survival analysis and independent prognostic analysis,and found that patients’ Risk Score and tumor pathological Stage could be used as independent prognostic factors.Then,we presented the prognostic risk of patients with a graph.At the same time,ROC curve and DCA decision curve were plotted with Calibration Plot to verify the model from three dimensions: differentiation,clinical applicability and calibration.Finally,in order to explore the specific regulatory relationships between variable clipping and shear factors,a regulatory network was constructed using Cytoscape software.