HUBioDataLab/CROssBARv2-KG

CROssBARv2-KG是一个异构的、通用的生物医学知识图谱数据集，包含了大约270万个节点和1260万条边，涵盖了14种不同的节点类型和51种不同的边类型。这些数据整合自34个生物数据源，并包含多个本体（如基因本体、Mondo疾病本体）以及丰富的节点和边属性。此外，数据集还生成了关键生物实体（如蛋白质、药物和基因本体术语）的嵌入，以增强语义深度。数据集分为节点和边两个主要目录，节点目录包含定义生物实体的CSV文件，边目录包含定义这些实体之间关系的CSV文件。

The CROssBARv2-KG is a heterogeneous and general-purpose biomedical knowledge graph dataset comprising approximately 2.7 million nodes and 12.6 million edges, spanning 14 distinct node types and 51 different edge types. These data are integrated from 34 biological data sources and include several ontologies (e.g., Gene Ontology, Mondo Disease Ontology) along with rich node and edge properties. Additionally, the dataset generates embeddings for key biological entities (e.g., proteins, drugs, and Gene Ontology terms) to enhance semantic depth. The dataset is organized into two primary directories: nodes, containing CSV files defining biological entities, and edges, containing CSV files defining relationships between these entities.