CROssBAR: comprehensive resource of biomedical relations with knowledge graph representations

Area of Science:

• Bioinformatics
• Systems Biology
• Computational Biology

Background:

• Large-scale biological and biomedical data are crucial for understanding disease mechanisms and developing treatments.
• Data fragmentation across resources and technologies hinders integrative multi-omics analysis.

Purpose of the Study:

• To develop a novel methodology for integrating and representing diverse biological data.
• To construct a comprehensive biological data resource, CROssBAR, to facilitate systems-level research.

Main Methods:

• Integrated large-scale biological/biomedical data from multiple resources into a NoSQL database.
• Employed deep learning to predict relationships between data entries, enriching the dataset.
• Developed interactive knowledge graphs for on-the-fly visualization of complex biological networks based on user queries.

Main Results:

• Created CROssBAR, a system integrating heterogeneous biological data, including genes, proteins, pathways, phenotypes, diseases, and drugs.
• Generated biologically meaningful modules through rigorous analysis of enriched data.
• Enabled easy-to-interpret, interactive knowledge graphs for systems-level research.

Conclusions:

• CROssBAR provides a comprehensive platform for integrative analysis of multi-omics data.
• The system facilitates the inference of biological mechanisms, gene-protein relationships, and disease associations.
• CROssBAR knowledge graphs are expected to advance systems-level research and therapeutic development.