山西医科大学学报2025，Vol.56Issue(4)：435-444,10.DOI:10.13753/j.issn.1007-6611.2025.04.013
基于语义三元组的尿毒症中药知识发现模型
Knowledge discovery model of traditional Chinese medicine for uremia based on semantic triplets
摘要
Abstract
Objective To propose a method of knowledge discovery of uremia-traditional Chinese medicine based on scientific re-search literature and biomedical database,and preliminarily verify the results,which provides ideas for exploring the biological mechanism of uremia and the treatment of traditional Chinese medicine.Methods The SemRep extraction tool was used to propose a disease-relationship-gene triad for uremia from the abstract text of the uremia research literature to obtain potential genes for uremia.The core genes of uremia were obtained by protein interaction network analysis of potential genes based on String database.GO analysis and pathway analysis were performed on the core genes to reveal the biomolecular mechanism of uremia.The Chinese medicines with regulatory effects on the core genes were screened from the Coremine Medical Chinese medicine database,and the Chinese medicine co-occurrence network was established and clustered to extract the core Chinese medicine components of each class group.Results A total of 137 potential genes for uremia were extracted from 25 439 publications,and 15 core genes,such as IL-6,IL-1B,CRP,IL-10,TGFB1,IGF-1,LEP,CXCL8,AGT,CCL2,TLR4,FGF2,IL-2,ACE and SERPINE1,were obtained by protein interaction network.In terms of biological processes,the core genes were involved in monocyte proliferation,etc.In terms of molecular functions,the core genes were enriched in receptor-ligand activity,etc.In cellular component,the core genes were involved in platelet alpha granule lumen,etc.The core genes were enriched in AGE-RAGE and IL-17 pathways,and so on.Potential Chinese medicines for ure-mia were classified into four class groups,which have effects on diabetic nephropathy,chronic renal failure,nephrotic syndrome,and chronic kidney disease,respectively.The Chinese medicines represented by ginseng,membranous astragalus,ginger bark,and ich-thyolimus were the core medicines corresponding to the four classes of diseases,respectively.Conclusion The drug knowledge dis-covery path constructed in this paper can effectively mine the knowledge from large-scale literature data on diseases and genes,occur-rence and development mechanism,and the mined knowledge can be verified in the literature.关键词
SPO/知识发现/尿毒症/知识抽取/中药预测/社会网络分析Key words
SPO/knowledge discovery/uremia/knowledge extraction/TCM prediction/social network analysis分类
社会科学引用本文复制引用
邰杨芳,安鹏伟,华国旻,昝彭..基于语义三元组的尿毒症中药知识发现模型[J].山西医科大学学报,2025,56(4):435-444,10.基金项目
国家社会科学基金资助项目(21BTQ050) （21BTQ050）
山西省高等学校一般性教学改革创新项目(J20230538) （J20230538）
山西省哲学社会科学规划(一般)课题基金项目(2021YJ117) （一般）
山西省研究生教育教学改革课题-研究生创新创业教育项目(2021YJJG114) （2021YJJG114）
