| Objective:To evaluate the predictive value of immune-related ferroptosis genes on the prognosis of patients with thyroid cancer(THCA)by using bioinformatics methods and to explore the mechanisms that affect the occurrence and development of THCA from multiple perspectives.Methods:1.Download clinical data and transcriptome sequencing data of THCA patients from TCGA and GEO databases,and download ferroptosis gene lists from FerrDb and immune-related gene lists from InnateDB databases.The THCA samples with complete clinical data were randomly divided into the training set and the validation set.2.After obtaining the crossed immune-related differentially expressed genes and ferroptosis differentially expressed genes,weighted gene co-expression network analysis(WGCNA),cluster analysis,univariate cox regression analysis,and coexpression analysis were used to screen out 9 prognoses,immune-related ferroptosis genes.3.Use Lasso regression analysis to obtain two genes,BID and CDKN2A,in the training set.Use the optimal penalty parameter(λ)to construct a multivariate Cox regression model and calculate its correlation coefficient.4.Obtain the corresponding risk score formula,calculate the median value of the training set,test set,and total central risk score respectively,and divide the patients into the high-risk group and the low-risk group.5.Use the receiver operating characteristic curve,Kaplan-Meier survival curve,and consistency index to compare with other similar models to verify the prediction accuracy,reliability,and superiority of the risk model.6.After integrating age,clinical-grade and risk score,use "rms" to construct a Nomogram and verify its performance.7.Use StarBase(v2.0)to predict the ceRNA regulatory network of risk model genes.8.Use the R package "GenVisR" to assess the samples for somatic mutations and perform copy number variation analysis on the samples.9.Explore the relationship between risk score and immunotherapy predictors such as Micro satellite Instability(MSI),Tumor Mutation Burden(TMB),Tumor Immune Dysfunction,and Rejection(TIDE)score.10.Explore the relationship between risk score and drug sensitivity.11.Use external data,clinical samples,and immunohistochemistry to verify the differential expression of related genes in cancer and adjacent tissues.Results:1.Two genes,BID and CDKN2A,were identified as prognostic factors and used to construct a risk scoring model.The overall survival rate of THCA patients in the low-risk group was higher than that in the high-risk group in the training set,test set,and total set(p<0.05).2.The excellent performance of the risk score has been verified and outperformed other models(AUC>0.65).3.The predicted regulatory network composed of CYTOR,has-miRNA-873-5p,and CDKN2A could affect prognosis(P<0.01).4.Single copy number deletion of CDKN2A was associated with up-regulation of CDKN2A expression and poor prognosis(P<0.001).5.Risk scores were correlated with immunotherapy predictors such as TMB and MSI(P<0.05),and there were differences between the high-risk group and the lowrisk group(P<0.05).6.The half-maximum inhibitory concentration of some drugs was correlated with risk score(P<0.05),and the difference between the high-risk group and the lowrisk group was statistically significant(P<0.05).7.Verified by external data,clinical samples,and immunohistochemistry,the differential expression of related genes in cancer and adjacent tissues is consistent with the research results.Conclusion:The model constructed by the prognosis-related immune-related ferroptosis differentially expressed genes screened in this study can effectively predict the prognosis and immunotherapy effect of THCA patients. |
PDF Full Text Request