抗癌药物基因组学知识表示模型构建

目的：通过构建抗癌药物基因组学知识表示模型，丰富药物、基因、疾病以及个性化用药等之间的语义关系，为临床医生精准用药、联合用药等提供参考依据，为药学科研人员开展新药研发、老药新用等研究提供理论支持，为癌症患者查询药物知识提供服务支撑。方法：对DrugBank、RxNorm、FDA药品说明书等药物基因组数据进行整合和抽取，在通用药物基因组学知识表示模型的基础上，涵盖药物、基因、疾病3个基本维度，并拓展个性化用药、药物副作用等知识维度，设计知识表示框架，确立相关实体类型，发现和定义实体与实体之间的语义关系。在构建好的抗癌药物基因组学知识表示模型的基础上，以黑色素瘤相关药物为例，对知识表示模型进行填充，并对实体和实体间语义关系进行可视化表达。结果：实现了药物、基因、疾病、个性化用药、药物副作用5类实体的概念抽取，定义了实体间的15种语义关系，构建了黑色素瘤药物基因组学相关的136个知识三元组。结论：面向抗癌药物精准用药的药物基因组学知识表示模型可以确定癌症用药与基因突变、人群、药物副作用之间的关联和推动药物基因组学知识在临床研究中的应用。

Establishment of pharmacogenomics knowledge representation model for anticancer drugs

Objective To study the semantic relationship among drugs, genes, diseases and individualized use of drugs by establishing the pharmacogenomics knowledge representation model for anticancer drugs. Methods A pharmacogenomics knowledge representation model was established by integrating and extracting pharmacogenomics data from the instructions of drugs covered in DrugBank, RxNorm and FDA based on the general pharmacogenomics knowledge representation models, in which drugs, genes and diseases were covered and individualized use of drugs and interaction of drugs were expanded. The semantic relationship between entities was discovered, defined and visually expressed with the drugs used in treatment of melanoma as an example. Results The concepts of drugs, genes, diseases, individualized use of drugs and interaction of drugs were extracted, 15 semantic relationships between them were defined, and a ternary system for 136 knowledge representations related with the pharmacogenomics of melanoma was established. Conclusion Pharmacogenomics knowledge representation model for precision use of anticancer drugs can identify the relationship of drugs used in treatment of cancer with gene definition, population, adverse effect of drugs and promote the application of pharmacogenomics knowledge in clinical studies.