Clinical trial information mapped to the OpenBioLink and custom KGs.

For each possible pairing of a drug-disease database (i.e., column 2) with entities that could be mapped to either the OpenBioLink or custom KG, we report the proportion of drug-disease pairs contained in ClinicalTrials.gov (i.e., column 3). We consider these clinically investigated drug-disease pairs as positive labels for the validation of our approach. The number of unique drugs (PubChem compound identifiers) and diseases (MONDO identifiers) are reported in columns 4 and 5, respectively, while the total number of possible combinations of these unique drugs and diseases are presented in column 6. (XLSX)

针对药物疾病数据库（即第2列）与可映射至OpenBioLink或自定义知识图谱（Knowledge Graph, KG）的实体的每一组可能配对，我们统计了ClinicalTrials.gov（即第3列）中收录的药物-疾病对占比。我们将这些经临床研究的药物-疾病对作为验证本研究方法的正样本标签。唯一药物（PubChem化合物标识符，PubChem compound identifiers）与唯一疾病（MONDO标识符，MONDO identifiers）的数量分别列于第4列和第5列，而上述唯一药物与疾病的全部可能组合总数则展示于第6列。（XLSX）