目的 基于专利数据挖掘与转录组学分析,探索中药复方治疗肌萎缩侧索硬化症(amyotrophic lateral sclerosis,ALS)的配伍规律及精准治疗策略。方法 从中国专利数据库筛选ALS中药复方专利,通过频次统计与关联分析挖掘核心药物及配伍规律;利用中药系统药理学数据库(traditional Chinese medicine systems pharmacology,TCMSP)和中药高通量实验和参考数据库(a high-throughput experiment and reference-guided database of TCM,HERB)构建复方-靶点网络,通过模块划分识别核心靶点模块;基于转录组数据集,采用非负矩阵分解(non-negative matrix factorization,NMF)对ALS患者进行转录组分型;运用网络相似性分析(Vertex/Edge Overlap,VEO算法)筛选各亚组优势复方,结合重启随机游走(random walk with restart,RWR)和SymMap平台预测证候特征。结果 通过对中国专利数据库中11项治疗ALS的中药复方专利(含101味中药)进行配伍规律分析,发现黄芪、茯苓、人参、淫羊藿和当归为高频核心药物。关联规则分析进一步揭示关键药对组合,其中白术-淫羊藿、白术-茯苓的配伍关联性最强。基于GSE16989数据集的转录组NMF分析,ALS患者被划分为4个亚组:组1显著富集白细胞介素-17(interleukin 17,IL-17)信号通路;组2未富集显著通路;组3与组4共同富集病毒-细胞因子互作通路。通过VEO算法计算复方蛋白质相互作用(protein-protein interaction,PPI)网络与各亚组的相似性,筛选出亚组特异性优势复方:复方11匹配组1;复方9匹配组2;复方8匹配组3、4。RWR算法结合SymMap平台进行亚组证候分型。结论 研究揭示了ALS的潜在优势中药复方,将转录组亚型与中医证候关联,为ALS“病证结合”精准治疗提供科学依据。

Objective To explore the compatibility patterns and precision treatment strategies of traditional Chinese medicine (TCM) prescriptions for amyotrophic lateral sclerosis (ALS) using patent data mining and transcriptomic analysis. Methods TCM prescriptions patents for ALS were screened from the Chinese Patent Database. Frequency statistics and association analysis were applied to identify core herbs and compatibility patterns. The traditional Chinese medicine systems pharmacology (TCMSP) and a high-throughput experiment- and reference-guided database of TCM (HERB) databases were utilized to construct prescriptions-target networks, and module partitioning was employed to identify core target modules. ALS patients were classified into transcriptional subgroups via non-negative matrix factorization (NMF) based on the GSE16989 transcriptomic dataset. Vertex/Edge Overlap (VEO) algorithm was used to screen subgroups-specific optimal prescriptions, while restart random walk (RWR) and the SymMap platform were applied to predict syndrome characteristics. Results Analysis of 11 ALS TCM prescriptions patents (containing 101 herbs) revealed Astragali Radix, Poria, Ginseng Radix et Rhizoma, Epimedii Folium, and Angelicae Sinensis Radix as high-frequency core herbs. Association rule mining identified key herb pairs, with Atractylodis Macrocephalae Rhizoma-Epimedii Folium and Atractylodis Macrocephalae Rhizoma-Poria showing the strongest compatibility correlations. NMF analysis classified ALS patients into four molecular subgroups: group 1 was significantly enriched in the interleukin-17 (IL-17) signaling pathway, group 2 showed no significant pathway enrichment; group 3 and group4 were co-enriched in the viral protein-cytokine receptor interaction pathway. Subgroups-specific optimal prescriptions were identified via VEO similarity: prescriptions 11 for group1, prescriptions 9 for group 2, and prescriptions 8 for group 3 and group 4. Syndrome stratification was performed using RWR and SymMap. Conclusion This study elucidates the precision treatment potential of TCM prescriptions for ALS by integrating transcriptomic subgroups with TCM syndrome characteristics, providing a scientific foundation for “disease-syndrome combination” therapy in ALS.

TP18;R285

国家自然科学基金面上项目（82474682）；国家“重大新药创制”科技重大专项（2017ZX09031059）