目的 基于生物信息学技术构建阿司匹林抵抗（aspirin resistance，AR）分子预测模型，探究复方丹参滴丸改善AR潜在机制。方法 在基因表达数据库（Gene Expression Omnibus，GEO）中获取AR和阿司匹林敏感（aspirin sensitivity，AS）相关的芯片数据，筛选差异表达基因（diffially express genes，DEGs），对AR和AS的DEGs进行富集分析，注释其功能；采用机器学习算法构建AR分子预测模型；在中国知网、万方、维普、PubMed、Web of Science数据库中获取复方丹参滴丸指纹图谱，根据Lipinski规则预测药物活性成分，使用Swiss Target Prediction数据库预测活性成分作用靶点，将DEGs和药物靶点取交集，获得复方丹参滴丸改善AR的潜在靶点，在metascape平台中对靶点基因进行功能注释；使用Autodock软件对核心靶点与活性成分进行分子对接；应用单样本基因集富集分析（single sample gene set enrichment analysis，ssGSEA）算法对复方丹参滴丸干预AR的每个靶点分别进行通路富集分析，最后构建复方丹参滴丸改善AR的“成分-靶点-通路-功能”桑基图。结果 共得到AR相关DEGs 597个，其中上调基因292个，下调基因305个。DEGs富集结果提示，AR相关基因主要定位在血小板α颗粒、分泌颗粒腔、细胞质囊泡腔上，通过介导花生四烯酸代谢、氧化磷酸化、WNT信号通路、白细胞介素-17（interleukin-17，IL-17）信号通路、类固醇同源物生物合成等信号通路，参与血液凝固、血小板活化、核因子-κB（nuclear factor-κB，NF-κB）转录因子活性、脂质代谢过程的负调控、白细胞活化相关炎症反应、肾上腺素受体信号通路、氧化应激反应等生物过程，发挥调控肝素结合、胰岛素受体结合、血小板衍生生长因子受体结合、肾上腺素受体活性、生长因子受体结合、磷脂酰肌醇激酶活性、主要组织相容性复合物II（MHC II）类受体活性等功能。基于机器学习算法获得9个AR的分子标志物，并构建了曲线下面积（area under curve，AUC）＝1的精准预测模型。通过网络药理学的方法获得复方丹参滴丸改善AR的10个活性成分和14个作用靶点及3个核心靶点，分子对接结果表明核心靶点和所有活性成分均能自发结合，富集分析提示复方丹参滴丸能够通过调控血栓形成、糖脂代谢、炎症反应、氧化应激、激素水平改善AR。结论 由9个AR的分子标志物组成的AR预测模型具有较为精准的预测性能；除血小板活化和聚集外，炎症反应、激素水平、氧化应激、糖脂代谢同样参与了AR的形成，复方丹参滴丸能够通过多靶点调控以上途径改善AR。

Objective To construct a molecular prediction model for aspirin resistance (AR) based on bioinformatics technology and to investigate the potential mechanism of Compound Danshen Dripping Pills (复方丹参滴丸) for reliving AR. Methods Microarray data of AR and aspirin sensitivity (AS) were obtained from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) of AR and AS were screened by R software, and their functions were annotated by enrichment analysis. AR molecular prediction models were constructed by machine learning methods. The fingerprint of Compound Danshen Dripping Pills were obtained from CNKI, Wanfang database, VIP, PubMed and Web of Science databases; The active ingredient was predicted according to Lipinski’s rule; The active ingredient targets were predicted using Swiss Target Prediction database; The DEGs and drug targets were intersected to obtain the potential targets of Compound Danshen Dripping Pills for relieving AR. The functional annotation of target genes was performed on the metascape platform; The molecular docking of core targets and active ingredients was performed using Autodock software; The single sample gene set enrichment analysis (ssGSEA) algorithm was applied to analyze the pathway enrichment of each target of Compound Danshen Dripping Pills for AR intervention, and finally the “component-target-pathway-function” sankey diagram of Compound Danshen Dripping Pills for AR relieving was constructed. Results A total of 597 AR-related DEGs were obtained, including 292 up-regulated genes and 305 down-regulated genes. The enrichment results of DEGs suggested that AR-related genes were mainly localized in platelet α granules, secretory granule lumen, cytoplasmic vesicle lumen, and were involved in many biological processes, such as blood coagulation, platelet activation, nuclear factor kappa-B (NF-κB) transcription factor activity, negative regulation of lipid metabolic process, leukocyte activation-related inflammatory response, adrenergic receptor signaling pathway, oxidative stress response, and etc, by mediating arachidonic acid metabolism, oxidative phosphorylation, WNT signaling pathway, interleukin-17 (IL-17) signaling pathway, steroid homologue biosynthesis and other signal pathways, thus playing a role in regulating heparin binding, insulin receptor binding, platelet-derived growth factor receptor binding, adrenergic receptor activity, growth factor receptor binding, phosphatidylinositol kinase activity, MHC class II receptor activity, and other functions. Nine molecular markers of AR were obtained based on machine learning algorithms, and an accurate prediction model with area under curve (AUC) = 1 was constructed. Ten active components, 14 action targets and three core targets of Compound Danshen Dropping Pills for improving AR were obtained through network pharmacology. The molecular docking results showed that the core targets and all active ingredients could bind spontaneously, and the enrichment analysis suggested that Compound Danshen Dripping Pills could improve AR by regulating thrombosis, glucolipid metabolism, inflammatory response, oxidative stress, and hormone levels. Conclusion The AR prediction model consisting of nine molecular markers of AR had accurate prediction performance.In addition to platelet activation and aggregation, inflammatory response, hormone levels, oxidative stress, and glucolipid metabolism might also be involved in the formation of AR, and Compound Danshen Dripping Pills can relieve AR through multi-targeted modulation of these pathways.

Q811.4;R285;TP18

天津市教委科研计划项目（2022ZD052）;国家自然科学基金青年基金项目（81803930）;国家自然科学基金项目（81873149）;国家“重大新药创制”科技重大专项资助项目（2018ZX09734002）;天津市“项目＋团队”重点培养专项：抗凝作用中西药对筛选系统研制（XC202033）