A knowledge graph of clinical trials ([Formula: see text])

Area of Science:

• Biomedical Informatics
• Clinical Research
• Data Science

Background:

• Clinical trials are crucial for drug development but are expensive and prone to failure.
• High costs and low success rates necessitate innovative approaches to leverage existing clinical trial data.
• Inferring knowledge from past trials can optimize the design of future clinical studies.

Purpose of the Study:

• To construct the first publicly available Clinical Trials Knowledge Graph (CTKG).
• To represent medical entities (studies, drugs, conditions) and their relationships within clinical trials.
• To demonstrate the utility of the CTKG for applications like drug repurposing and similarity search.

Main Methods:

• Developed a novel knowledge graph structure to capture clinical trial information.
• Integrated diverse data sources to populate the knowledge graph with nodes and edges.
• Performed embedding analysis on the constructed knowledge graph.

Main Results:

• Successfully constructed the Clinical Trials Knowledge Graph (CTKG).
• The CTKG effectively represents complex relationships between medical entities in clinical trials.
• Embedding analysis revealed the graph's potential for practical applications.

Conclusions:

• The Clinical Trials Knowledge Graph (CTKG) offers a valuable resource for the research community.
• The CTKG can facilitate more efficient and effective clinical trial design.
• The demonstrated applications highlight the potential of the CTKG for advancing drug discovery and repurposing.