Reactome has released two preprints describing recent advances in applying artificial intelligence to pathway curation and user access.

The preprint, Application of Large Language Models for Annotating Genes into Reactome Pathways, presents an LLM-assisted workflow that supports expert curators by identifying candidate pathways for genes, retrieving relevant literature, and generating mechanistic summaries. Evaluation shows that the approach can meaningfully support manual curation while preserving curator oversight.

Screenshots of the LLM App in the web-based Reactome curation tool for query gene input, configuration and annotated pathways.

The preprint, React-to-Me: A Conversational Interface for Interactive Exploration of the Reactome Pathway Knowledgebase, introduces a grounded conversational assistant that enables natural language queries over Reactome content. The system enforces source traceability to curated data and avoids speculative responses, improving accessibility without sacrificing factual accuracy.

Graphical abstract showing React-to-Me. 

Together, these studies outline a practical framework for integrating AI into Reactome to enhance scalability and usability while maintaining scientific rigor.