基于SPO语义三元组的自闭症谱系障碍药物知识发现

打开文本图片集

Drug knowledge discovery for autism spectrum disorders based on SPO predications

LYU Yanhua， ZHAO Hongxia， LI Qi， LIANG Aoxue， YU Qi

School of Management， Shanxi Medical University， Shanxi 030001 China

Corresponding Author  LYU Yanhua， E⁃mail：lvyanhua01@163.com

Abstract  Objective：To extract SPO（Subject⁃Predicate⁃Object，SPO） from literature related to Autism Spectrum Disorders（ASD） using semantic mining technology and construct a knowledge graph of ASD drug entities，to explore the potential drug for the treatment of ASD at a deeper level， and provide new ideas for discovering valuable potential drugs for other diseases（https：//clinicaltrials.gov）.Methods：Using the tools SemRep and Metamap based on the Unified Medical Language System （ULMS） to process ASD literature records and obtain SPO of ASD drug entities.The Neo4j database was used for knowledge storage to  construct an ASD drug entities knowledge graph.Using three semantic pathways to discovery ASD drug knowledge based on the knowledge graph.Then verified and analyzed the effectiveness of the results in the clinical trials databases.Results：The SPO obtained includes 1 262 head entities， 687 tail entities， and 18 entity relationships.A total of 32 drugs were discovered through three semantic pathways，27 potential drugs for ASD was screened out，and 19 drugs can be validated in the clinical trials databases.Conclusions：The knowledge discovery of ASD drugs based on knowledge graph which built by SPO can provide a certain theoretical and methodological basis for drug repositioning，provide new ideas for traditional drug discovery，and provide decision support for clinical experiments and scientific research.

Keywords    autism spectrum disorders； knowledge graph； semantic mining； drug repositioning

摘要  目的：运用语义挖掘技术抽取自闭症相关文献中的三元组并构建自闭症药物实体知识图谱，深层次开展自闭症治疗的潜力药物知识发现，同时也为其他疾病发现有价值的潜在治疗药物提供新思路。（剩余13353字）