构建可语义推理的药物效应知识图谱，助力药物重定位与安全监测。
SIDEKICK: A Semantically Integrated Resource for Drug Effects, Indications, and Contraindications
- 用大模型提取药监局标签，映射到疾病与表型本体
- 覆盖5万+药品说明书，比SIDER、ONSIDES更优用于副作用相似性重定位
- 支持自动安全监控，适合生物医学研究与临床决策系统
药物警戒和临床决策支持系统依赖结构化药物安全数据。然而，现有数据集多基于MedDRA等术语体系，限制了语义推理能力及与语义网本体和知识图谱的互操作性。为此，我们构建了SIDEKICK知识图谱，从FDA结构化产品标签中标准化药物适应症、禁忌症和不良反应。采用基于大语言模型（LLM）的提取与图检索增强生成（Graph RAG）的工作流，将超过5万份药品说明书中的术语映射至人类表型本体（HPO）、MONDO疾病本体和RxNorm。该资源在基于副作用相似性的药物重定位任务中优于SIDER和ONSIDES数据库。我们以资源描述框架（RDF）形式序列化数据，并使用语义科学集成本体（SIO）作为上层本体，进一步提升互操作性。由此，SIDEKICK支持自动化安全监测与基于表型的相似性分析，推动药物重定位研究。
原文摘要 · Abstract (English)
Pharmacovigilance and clinical decision support systems utilize structured drug safety data to guide medical practice. However, existing datasets frequently depend on terminologies such as MedDRA, which limits their semantic reasoning capabilities and their interoperability with Semantic Web ontologies and knowledge graphs. To address this gap, we developed SIDEKICK, a knowledge graph that standardizes drug indications, contraindications, and adverse reactions from FDA Structured Product Labels. We developed and used a workflow based on Large Language Model (LLM) extraction and Graph-Retrieval Augmented Generation (Graph RAG) for ontology mapping. We processed over 50,000 drug labels and mapped terms to the Human Phenotype Ontology (HPO), the MONDO Disease Ontology, and RxNorm. Our semantically integrated resource outperforms the SIDER and ONSIDES databases when applied to the task of drug repurposing by side effect similarity. We serialized the dataset as a Resource Description Framework (RDF) graph and employed the Semanticscience Integrated Ontology (SIO) as upper level ontology to further improve interoperability. Consequently, SIDEKICK enables automated safety surveillance and phenotype-based similarity analysis for drug repurposing.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。