Abstract
Emerging proof shows that abnormal lipometabolism affects invasion, metastasis, stemness and tumor microenvironment in carcinoma cells. However, molecular markers related to lipometabolism have not been further established in breast cancer. In addition, numerous studies have been conducted to screen for prognostic features of breast cancer only with RNA sequencing profiles. Currently, there is no comprehensive analysis of multiomics data to extract better biomarkers. Therefore, we have downloaded the transcriptome, single nucleotide mutation and copy number variation dataset for breast cancer from the TCGA database, and constructed a riskScore of twelve genes by LASSO regression analysis. Patients with breast cancer were categorized into high and low risk groups based on the median riskScore. The high-risk group had a worse prognosis than the low-risk group. Next, we have observed the mutated frequencies and the copy number variation frequencies of twelve lipid metabolism related genes LMRGs and analyzed the association of copy number variation and riskScore with OS. Meanwhile, the ESTIMATE and CIBERSORT algorithms assessed tumor immune fraction and degree of immune cell infiltration. In immunotherapy, it is found that high-risk patients have better efficacy in TCIA analysis and the TIDE algorithm. Furthermore, the effectiveness of six common chemotherapy drugs was estimated. At last, high-risk patients were estimated to be sensitive to six chemotherapeutic agents and six small molecule drug candidates. Together, LMRGs could be utilized as a de novo tumor biomarker to anticipate better the prognosis of breast cancer patients and the therapeutic efficacy of immunotherapy and chemotherapy.