Revealing Cell Division-Related Gene as a Potential Diagnostic Gene Biomarker of Triple-Negative Breast Cancer Based on Machine Learning Analysis

He J, Ding HL, Zhou B, Huang C and Ouyang QW

Yi Chun People’s Hospital, Yi Chun, China
Nanchang People’s Hospital, Jiangxi Province Key Laboratory of Breast Diseases, China

^*Correspondance to: Qian-Wen Ouyan 

 PDF  Full Text Research Article | Open Access

Abstract:

Objective: To identify diagnosis markers that can differentiate Triple-Negative Breast Cancer (TNBC) from non-TNBC in breast cancer patients through extensive gene expression data analysis. Methods: Use the analysis of expression profiling and clinical data of TNBC patients sourced from the Gene Expression Omnibus (GEO) databases. The 'limma' R package was initially used to identify Differentially Expressed Genes (DEGs) associated with diagnosis by comparing tumor and normal tissues. Subsequent analyses included enrichment analysis and Protein-Protein Interaction (PPI) analysis. Gene modules associated with TNBC were obtained using Weighted Gene Co-expression Network Analysis (WGCNA). Least absolute shrinkage and selection operator (Lasso) based on the results of WGCNA and CytoHubba analysis. A ROC curve was then generated to assess the polygenic risk scores in individuals with TNBC. Results: The analysis of the GSE76275 GEO dataset revealed 324 DEGs, with 128 being upregulated and 196 downregulated. CytoHubba's five algorithms were then utilized to identify the top 30 hub genes on the PPI network map, resulting in the identification of 7 key genes through intersection. The entire dataset underwent analysis using the WGCNA package, which identified a yellow module (r=0.4, P<0.001) that showed a significant correlation with morbidity. The intersection of 7 hub genes identified through gene and CytoHubba screening resulted in 6 key genes. Enrichment analysis using the clusterProfiler package revealed their involvement associated with cell division. Subsequent analysis with the pROC package showed that each gene had a good diagnostic efficacy, as indicated by the Area Under the Curve (AUC), in distinguishing between TNBC and non-TNBC cases. Conclusion: The comprehensive analysis of the GSE76275 GEO dataset, which included differential analysis, WGCNA, CytoHubba screening, and Lasso successfully identified and enriched key genes linked to the differences between TNBC and non-TNBC cases, with the cell division-related gene scoring high. Additionally, ROC analysis confirmed the diagnostic efficacy of these important genes in distinguishing between the two types.

Keywords:

Triple-negative breast cancer; Bioinformatics analysis; Differentially expressed genes

Cite the Article:

He J, Ding HL, Zhou B, Huang C, Ouyang QW. Revealing Cell Division- Related Gene as a Potential Diagnostic Gene Biomarker of Triple-Negative Breast Cancer Based on Machine Learning Analysis. Clin Oncol. 2024;9:2088..