PrimeKG

Here, we present the Precision Medicine Knowledge Graph (PrimeKG). This resource provides a holistic view of diseases. We have integrated 20 high-quality datasets, biorepositories and ontologies to curate this knowledge graph. PrimeKG systematically captures information about 17,080 diseases with 4,050,249 relationships representing various major biological scales, including diseases, drugs, genes, proteins, exposures, phenotypes, drug side effects, molecular functions, cellular components, biological processes, anatomical regions, and pathways. Disease nodes in our multi-relational knowledge graph are densely connected to every other node type. PrimeKG's rich graph structure is supplemented with textual descriptions of clinical guidelines for drug and disease nodes to enable multi-modal disease exploration. To get started with using PrimeKG, please explore our project website: https://zitniklab.hms.harvard.edu/projects/PrimeKG/

本研究提出精准医学知识图谱（Precision Medicine Knowledge Graph, PrimeKG）。该资源可为疾病领域提供全景式认知框架。本研究整合了20项高质量数据集、生物样本库与本体知识库，完成了该知识图谱的编汇构建。PrimeKG可系统性收录17080种疾病的相关信息，包含4050249条关联关系，覆盖疾病、药物、基因、蛋白质、暴露因素、表型、药物不良反应、分子功能、细胞组分、生物过程、解剖部位以及通路等多个核心生物学范畴。在该多关系知识图谱中，疾病节点与其余所有类型的节点均呈现高密度连接。PrimeKG丰富的图谱结构还补充了药物与疾病节点对应的临床指南文本描述，以支持多模态疾病探索研究。如需使用PrimeKG开展相关研究，请访问本项目的官方网站：https://zitniklab.hms.harvard.edu/projects/PrimeKG/

搜集汇总

背景与挑战

背景概述

PrimeKG是一个全面的精准医学知识图谱，整合了20个高质量数据源，覆盖17,080种疾病和超过400万种关系，涉及多种生物尺度。其特色在于提供了丰富的疾病-药物关联信息和临床指南文本，支持多模态疾病研究。