Storage and Query of Drug Knowledge Graphs Using Distributed Graph Databases: A Case Study.

Area of Science:

• Bioinformatics
• Computational Chemistry
• Drug Discovery

Background:

• Large-scale drug knowledge graphs are crucial for drug research, but their storage and querying present challenges.
• Distributed graph databases offer a promising solution for managing complex pathway queries.
• A research gap exists in evaluating the performance of these databases for drug knowledge graphs.

Purpose of the Study:

• To evaluate the feasibility and performance of distributed graph databases for managing large-scale drug knowledge graphs.
• To assess the storage and query performance of a drug knowledge graph system built on Nebula Graph.
• To validate the system's utility in drug repurposing using established benchmarks.

Main Methods:

• Designed a drug knowledge graph storage and query system utilizing the Nebula Graph database.
• Evaluated the system's performance in terms of data writing and query execution.
• Employed two drug repurposing benchmarks (repoDB and DrugRepoBank) for comprehensive assessment.

Main Results:

• Distributed graph databases demonstrated superior performance over single-machine databases in data writing, regular queries, constrained queries, and concurrent queries.
• Writing performance advantages increased with data volume, and query performance benefits grew with task complexity.
• Drug repurposing benchmarks identified 78.54% of pathways consistent with approved treatments and 12 potential new drug indications with literature support.

Conclusions:

• The developed system effectively constructs, stores, and queries extensive multisource drug knowledge graphs.
• The system provides reliable and explainable drug-disease paths, supporting drug repurposing initiatives.
• Distributed graph databases are a viable and high-performing solution for large-scale drug knowledge graph management.