Screening of Key Genes for Triple-Negative Breast Cancer Based on a Differential Co-expression Method

The screening of key genes in triple-negative breast cancer is mainly studied in this paper. A diagnostic model based on the differential co-expression network method DiffCoEx is investigated. The result of commonly used differential coexpression analysis are modules composed of a large number of genes, with no further modeling methods provided to narrow down the gene modules for deeper exploratory analysis. The purpose of this study is to establish a diagnostic model that combines differential co-expression network and Lasso regression to screen key genes of triple-negative breast cancer. Data sets were collected including 36 cases of triple-negative breast cancer and 36 cases of non-triple-negative breast cancer. Five key genes of triple-negative breast cancer were obtained by using the adjacency matrix, topological overlap, hierarchical clustering, and sequencing test constructed by the differential co-expression network analysis DiffCoEx and modeling and screening by the Lasso regression. The reliability and validity of the model and index genes are demonstrated by the area under the subject working characteristic curve of the verification set of 67 samples.