Breast cancer (BC) is the most malignant tumor in women. The molecular mechanisms underlying tumorigenesis still need to be further elucidated. It is necessary to investigate novel candidate genes involved in breast cancer progression and prognosis. In this study, we commit to explore candidate genes that associate with prognosis and therapy in BC by a comprehensive bioinformatic analysis. Four GEO datasets (GSE5764, GSE7904, GSE20711, and GSE29431) and the BC-related transcriptome data in TCGA database were downloaded and used to identify the differently expressed genes (DEGs). The function of DEGs was analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment analysis. The protein–protein interaction (PPI) network of DEGs was constructed to identify hub genes. Prognostic candidate genes were identified through survival analysis. In addition, potential therapeutic targets were identified by constructed gene–drug interaction network through Comparative Toxicogenomics Database. A total of 547 DEGs (302 up and 245 down) were identified. Three core-subnetwork and 25 hub genes were identified in PPI network. Seven genes (namely COL12A1, QPRT, MRPL13, KRT14, KRT15, LAMB3, and MYBPC1) were identified as crucial prognostic candidate genes, which significantly associated with breast cancer overall survival. Furthermore, two representative candidate genes (COL12A1 and LAMB3) were optionally chosen for verification by reverse transcription and quantitative real-time polymerase chain reaction (RT-PCR). What's more, the gene–drugs interaction analysis indicates several antitumor drugs that could affect the expression of these prognostic markers, such as doxorubicin, cisplatin, and tamoxifen. These results identified seven crucial candidate genes that may serve as prognosis biomarkers and novel therapeutic targets of breast cancer, which may facilitate further understanding the molecular pathogenesis and providing potential therapeutic strategies for BC.

中文翻译：

---

与乳腺癌预后相关的候选基因的鉴定。

乳腺癌（BC）是女性中最恶性的肿瘤。肿瘤发生的分子机制仍需进一步阐明。有必要研究涉及乳腺癌进展和预后的新候选基因。在这项研究中，我们致力于通过全面的生物信息学分析探索与BC预后和治疗相关的候选基因。TCGA数据库中的四个GEO数据集（GSE5764，GSE7904，GSE20711和GSE29431）和与BC相关的转录组数据已下载并用于鉴定差异表达的基因（DEG）。通过基因本体论（GO）和《京都基因与基因组百科全书》（KEGG）途径富集分析来分析DEG的功能。构建了DEG的蛋白质间相互作用（PPI）网络，以识别轮毂基因。通过生存分析确定了预后候选基因。此外，通过比较毒物基因组数据库通过构建的基因-药物相互作用网络确定了潜在的治疗靶标。总共确定了547个DEG（302个向上和245个向下）。在PPI网络中鉴定出3个核心子网络和25个集线器基因。七个基因（即COL12A1，QPRT，MRPL13，KRT14，KRT15，LAMB3和MYBPC1被鉴定为至关重要的预后候选基因，这些基因与乳腺癌的总体生存率显着相关。此外，两个代表性候选基因（COL12A1和LAMB3通过反转录和定量实时聚合酶链反应（RT-PCR）进行验证）。此外，基因-药物相互作用分析表明，几种抗肿瘤药物可能会影响这些预后标志物的表达，例如阿霉素，顺铂和他莫昔芬。这些结果确定了七个至关重要的候选基因，它们可以作为乳腺癌的预后生物标志物和新的治疗靶标，从而有助于进一步了解分子发病机制，并为BC提供潜在的治疗策略。