血管新生对肿瘤的进展、侵袭性行为和转移起着贡献。尽管一些内皮功能异常基因（血管新生相关基因 [ARG]）已被确定为乳腺癌诊断生物标志物，但ARG的混合效应尚未得到彻底研究。获得乳腺癌的RNA测序数据和患者生存数据以进行进一步分析。MSigDB网站包括血管新生相关机制。一致性聚类分析将1082名乳腺癌患者分为三个簇。使用limma包识别差异表达基因（DEGs）。结合GO和基因集富集分析（GSEA）来识别两个预定义簇之间的细胞遗传学功能。然后选择Serpin家族F成员1（SERPINF1）、angiomotin（AMOT）、坏血酸协同哺乳细胞白血病基因（PML）和BTG抗增殖因子1（BTG）来构建随机森林生存分析的预测模型。使用GSE58812三阴性乳腺癌队列作为验证集进行外部验证。中值评分系统用于识别高风险和低风险组，其诊断结果有显著差异。使用单样本基因集富集分析（ssGSEA）和xCell算法计算免疫浸润评分，使用R软件包"oncoPredict"计算Genomics of Drug Sensitivity in Cancer（GDSC）数据库中药物的敏感性评分。此外，使用来自GSE118389的单细胞转录组测序信息对七例三阴性乳腺癌进行单细胞分析，阐明SERPINF1、AMOT、PML和BTG1的解释。总之，本研究构建了以ARG为中心的疾病模型，不仅可以预测治疗药物，还可以预测其机制轨迹，从而为乳腺癌诊断的不同患者群体提供个体化治疗建议。© 2023 Wiley Periodicals LLC.

Angiogenesis contributes to tumor progression, aggressive behavior, and metastasis. Although several endothelial dysfunction genes (angiogenesis-related genes [ARGs]) have been identified as diagnostic biomarkers of breast cancer in a few studies, the mixed effects of ARGs have not been thoroughly investigated. The RNA sequencing data and patient survival datasets of breast cancer were obtained for further analysis. MSigDB website includes angiogenesis-related mechanisms. The consensus clustering analysis identifies 1082 breast cancer patients as three clusters. differential expression genes (DEGs) were identified by limma package. GO combined with gene set enrichment analysis (GSEA) to identify cytogenetic functions between two predefined clusters. Then Serpin Family F Member 1 (SERPINF1), angiomotin (AMOT), promyelocytic leukemia (PML), and BTG anti-proliferation factor 1 (BTG) were selected to construct prediction models using random forest survival analysis. External validation was performed using the GSE58812 triple-negative breast cancer cohort as the validation set. The median scoring system was used to discern the high- and low-risk groups, and there was a significant difference in their diagnostic results. Immunological infiltration scores were calculated using single sample gene set enrichment analysis (ssGSEA) and xCell algorithms, and consciousness scores were calculated using the R package "oncoPredict" for drugs in the Genomics of Drug Sensitivity in Cancer (GDSC) database. In addition, the single-cell analysis of seven triple-negative breast cancers using scRNA-seq information from GSE118389 demonstrated the interpretation of SERPINF1, AMOT, PML, and BTG1. In conclusion, this investigation engineered ARG-centric disease paradigms that not only prognosticated prospective therapeutic compounds, but also projected their mechanistic trajectories, thereby facilitating the proposition of tailored treatments within diverse patient cohorts diagnosed with breast cancer.© 2023 Wiley Periodicals LLC.