2.2. Construction and assessment of BRCA-related CDGs

First, a Pearson's correlation analysis was performed to select CDGs. We included only those CDGs with | Pearson R | >0.5 and p < 0.05 to ensure a robust analysis, thus, we found 5 CDGs that could meet the criteria. Univariate and multivariate Cox regression analyses were applied to the training and validation cohorts to determine the significance of the five prognostic CDGs. The coefficients of the five CDGs were obtained via cross-validation using the R package “glmnet” [40]. Then, the risk score was calculated as follows: risk score = (0.0175 × PIK3CA expression level) + (0.0254 × AIFM1 expression level) + (0.0195 × GABARAPL2 expression level) + (0.1076 × ATG4A expression level) - (0.00996 × NFKBIA expression level). We calculated the risk score for each patient in TCGA cohort. Subsequently, the patients were divided into low- and high-risk subgroups based on median risk scores. {"type":"entrez-geo","attrs":{"text":"GSE1456","term_id":"1456"}}GSE1456, available from the Gene-Expression Omnibus (GEO) database (https://www.ncbi.nlm. nih.gov/geo/), involving 159 cases, were used for external validation. {"type":"entrez-geo","attrs":{"text":"GSE7390","term_id":"7390"}}GSE7390, {"type":"entrez-geo","attrs":{"text":"GSE16446","term_id":"16446"}}GSE16446, {"type":"entrez-geo","attrs":{"text":"GSE20685","term_id":"20685"}}GSE20685, {"type":"entrez-geo","attrs":{"text":"GSE20711","term_id":"20711"}}GSE20711, {"type":"entrez-geo","attrs":{"text":"GSE42568","term_id":"42568"}}GSE42568, {"type":"entrez-geo","attrs":{"text":"GSE45255","term_id":"45255"}}GSE45255, {"type":"entrez-geo","attrs":{"text":"GSE4839","term_id":"4839"}}GSE4839, and {"type":"entrez-geo","attrs":{"text":"GSE58812","term_id":"58812"}}GSE58812 were obtained from the GEO database.