基于表型术语及CBK理念构建罕见病知识库

Construction of Rare Disease Knowledge Base Based on Phenotypic Terms and CBK Concept

通过扩展人类表型术语集(HPO)的中文同义词,通过自然语言处理程序对电子病历中的表型概念进行识别后,借助可计算生物医学知识(CBK)理论,重新构建了整合电子病历及权威知识库的罕见病诊断知识库。因真实世界罕见病患者信息还包括各种合并症及并发症,其疾病-表型关联更加混杂,整合二者知识库是对目前罕见病诊断知识库在国内EMR落地的有益探索。未来通过CBK及其KGrid平台,该罕见病诊疗知识可以进行聚合计算、推演,封装后以便提供知识即服务(KaaS)的云计算服务,最终提升知识的复用及互操作性。

In this study,the Chinese synonym of Human Phenotype Ontology(HPO)was enriched,and the phenotype concept in EMR was identified by natural language processing program.After that,the rare disease knowledge base integrating EMR and authoritative source was rebuilt by using the Computable Biomedical Knowledge(CBK)theory.For rare disease patients in the real world include various complications,the phenotype-disease associations are more mixed.The knowledge integration of EMR and authoritative source is a beneficial exploration and practice for knowledge application in Chinese EMR.In the future,through CBK and its KGrid platform,the knowledge base of rare diseases could be aggregated and inferred.The knowledge also could be packaged to provide a KaaS(knowledge as a service),a cloud computing services,and ultimately improve the reuse and interoperability of knowledge.