Abstract
BACKGROUND: Brown adipocytes are vital in cancer's emergence and progression. However, such significance of brown adipocytes-related genes (BARGs) in colorectal cancer (CRC) still awaits exploration. The potential prognostic mechanisms of BARGs in CRC were aimed to be explored in this study. METHODS: Datasets related to CRC were obtained from public databases. Candidate genes were pinpointed by synthesizing the findings from analysis of differential expression and Weighted Gene Co-expression Network Analysis (WGCNA). Utilizing Mendelian randomization (MR), sensitivity analysis, and the Steiger test, the feature genes causally linked to CRC were precisely pinpointed. Then, prognostic genes were identified using univariate Cox analysis and machine learning algorithms, and a prognostic model was developed and validated. In addition, the construction and evaluation of nomograms, analysis of the immune microenvironment, drug sensitivity, and functional enrichment were also carried out. RESULTS: Altogether, 168 characteristic genes were pinpointed through MR analysis. ADD2, PDE1B, LZTS1, and PCSK5 were identified as prognostic genes. The area under the curve (AUC) values of receiver operating characteristic (ROC) for both the prognostic model and the nomogram were greater than 0.6, indicating high accuracy. Additionally, positive correlations were found among the differentially expressed immune cells, and PDE1B exhibited the most significant positive correlation with regulatory T cells. A variety of chemotherapeutic drugs were more significantly effective in treating CRC patients in low-risk group (LRG). Gene Set Enrichment Analysis (GSEA) demonstrated that the pathways linked to cell microenvironment remodeling and migration regulation were significantly enriched between high-risk group (HRG) and LRG. CONCLUSION: ADD2, PDE1B, LZTS1, and PCSK5 were identified as prognostic genes. A prognostic model related to BARGs in CRC was established, offering fresh perspectives on the prognostic treatment of CRC.