Machine learning algorithm to construct cuproptosis- and immune-related prognosis prediction model for colon cancer

Integrated bioinformatics and experiment revealed that cuproptosis is the potential common pathogenesis of three kinds of primary cardiomyopathy

Cuproptosis is a recently reported new mode of programmed cell death which might be a potential co-pathogenesis of three kinds of primary cardiomyopathy. However, no investigation has reported a clear relevance between primary cardiomyopathy and cuproptosis. In this study, the differential cuproptosis-related genes (CRGs) shared by three kinds of primary cardiomyopathy were identified in training sets. As a result, four CRGs shared by three kinds of primary cardiomyopathy were acquired and they were mainly related to biological processes such as cell death and immuno-inflammatory response through differential analysis, correlation analysis, GSEA, GSVA and immune cell infiltration analysis. Then, three key CRGs (K-CRGs) with high diagnostic value were identified by LASSO regression. The results of nomogram, machine learning, ROC analysis, calibration curve and decision curve indicated that the K-CRGs exhibited outstanding performance in the diagnosis of three kinds of primary cardiomyopathy. After that, in each disease, two molecular subtypes clusters were distinguished. There were many differences between different clusters in the biological processes associated with cell death and immunoinflammation and K-CRGs had excellent molecular subtype identification efficacy. Eventually, results from validation datasets and in vitro experiments verified the role of K-CRGs in diagnosis of primary cardiomyopathy, identification of primary cardiomyopathic molecular subtypes and pathogenesis of cuproptosis. In conclusion, this study found that cuproptosis might be the potential common pathogenesis of three kinds of primary cardiomyopathy and K-CRGs might be promising biomarkers for the diagnosis and molecular subtypes identification of primary cardiomyopathy.

Evaluation the role of cuproptosis-related genes in the pathogenesis, diagnosis and molecular subtypes identification of atherosclerosis

Background

At present, the pathogenesis of atherosclerosis has not been fully elucidated, and the diagnosis and treatment face great challenges. Cuproptosis is a novel cell death pattern that might be involved in the development of atherosclerosis. However, no research has reported the correlation between cuproptosis and atherosclerosis.

Methods

The differential cuproptosis-related genes (CRGs) between atherosclerosis group and control group (A-CRGs) were discovered via differential expression analysis. The correlation analysis, PPI network analysis, GO, KEGG and GSEA analysis were performed to investigate the function of A-CRGs. The differences of biological function between atherosclerosis group and control group were investigated via immune infiltration analysis and GSVA. The LASSO regression, nomogram and machine learning models were constructed to predict atherosclerosis risk. The atherosclerosis molecular subtypes clusters were discovered via unsupervised cluster analysis. Subsequently, we used the above research methods to analyze the differential CRGs between clusters (M-CRGs) and evaluate the molecular subtypes identification performance of M-CRGs. Finally, we verified the diagnostic value for atherosclerosis and role in cuproptosis of these CRGs through the validation set and in vitro experiments.

Results

Five A-CRGs were identified and they were mainly related to the biological function of copper ion metabolism and immune inflammatory response. The diagnostic models and nomogram of atherosclerosis based on 5 A-CRGs indicated that these genes had well diagnostic value. A total of two molecular subtypes clusters were obtained in the atherosclerosis group. There were many differences in biological functions between these two molecular subtypes clusters, such as mitochondrial outer membrane permeabilization and primary immunodeficiency. In addition, 3 M-CRGs were identified in the 2 clusters. Machine learning models and nomogram constructed based on M-CRGs showed that these genes had well molecular subtypes identification efficacy. In the end, the results of in vitro experiment and validation set confirmed the diagnostic value for atherosclerosis and role in cuproptosis of these genes.

Conclusion

The cuproptosis may be a potential pathogenesis of atherosclerosis and CRGs may be promising markers for the diagnosis and molecular subtypes identification of atherosclerosis.