Review Article

A Review on the Application of Knowledge Graph Technology in the Medical Field

Table 2

Commonly used data sets for medical entity relation extraction.

NameDetails

DrugBankMore than 80 aspects are provided for each drug, including brand name, chemical structure, protein and DNA sequences, related links on the Internet, feature descriptions and detailed pathological information, and so on.

STITCHA platform for searching known and predicted interactions between compounds and proteins. The STITCH database contains more than 30,000 small molecular compounds and 2.6 million protein interactions from 1,133 species.

TCMSPIt includes 499 traditional Chinese medicine registered in Chinese Pharmacopoeia, including 29,384 ingredients, 3,311 targets, and 837 related diseases. This information can be queried and downloaded into the database. The disease information in this database comes from the TTD and PharmGKB databases.

TTDProvide information about drugs, targets, diseases, and pathways. The current version collects 34,019 drugs, including 2,544 licensed drugs, 8,103 clinical trial drugs, and 18,923 drugs under development. Each drug provides information on its chemical structure, targets, targeted diseases, and related pathways. Users can search the database through targets, medications, conditions, and biomarkers and use drug similarity search tools to predict the targets of compounds without target information.

CCHMCThe data are from CCHMC (Cincinnati Children’s Hospital Medical Center). CCHMC’s institutional review committee approved the release of the data. All outpatient chest x-ray films and revisit chest films were sampled for one year by the bootstrap method. These data are commonly used data, and they are designed to provide sufficient code to cover the actual proportion of pediatric radiology activities.

MIMICA publicly available data set developed by the Computational Physiology Laboratory of the Massachusetts Institute of Technology, including unidentified health data (including demographics, vital signs, laboratory tests, medication, etc.) are related to about 40,000 intensive care patients.