Scalable Curation of TF-TG interactions via Integration of NLP and VSM (2022-23)

Documented regulatory interactions between Transcription Factors (TFs) and target genes (TGs) form a crucial resource for biological network building. The ExTRI text mining project yielded about 54,000 sentences from literature that proposedly describe a TF-TG interaction, linked to database identifiers. This corpus is now a compelling target for a community curation effort, which will be performed by dedicated curators in this project, and modelers potentially contributing via Cytoscape. To make this large-scale effort possible, we will combine the intuitive and flexible curation tool VSM (Visual Syntax Method), with automated NLP (Natural Language Processing) assistance, into a community curation app. Premarked ExTRI sentences will be shown in their abstract, along with NLP annotations provided by Europe-PMC's API, to help curators formalize knowledge into a form suitable for computational analysis. The new web app will be designed for reusability, guided by professional curation partners, and evaluated for effect on curator performance.