Genomenon, a leading genomic intelligence company, has announced the creation of the Genomenon Genomic Graph (G³) knowledgebase, a comprehensive, searchable database of all clinically relevant, empirical genomic evidence. This marks the first time that information from the entire body of clinically relevant literature has been captured in a single, searchable resource. The G³ knowledgebase aims to transform how researchers and clinicians access and utilize genomic data, potentially accelerating discoveries and advancements in patient care.
The G³ knowledgebase is built upon an advanced AI platform that includes a genomic-specific large language model (LLM). This LLM was trained using Genomenon’s own expertly curated genomic datasets. This approach ensures a high level of accuracy and allows the system to handle the complexities of genomic information, including historical trends, diverse natural language descriptions, and evolving formal nomenclatures.

The knowledgebase includes a wide array of information, such as:

• Genes
• Genetic variants
• Copy number variants
• Structural variants (including fusion events)
• Gene-disease relationships
• Drugs
• Phenotypes
• Patient demographics and symptoms

The G3 knowledgebase allows for advanced biomedical literature searches using natural language queries. This enables users to prioritize, annotate, and summarize relevant articles more efficiently. The platform is also capable of generating actionable patient and disease insights and can produce interactive databases with structured generative AI insights and analytics, which are particularly useful for real-world evidence applications.

Source.