Monarch representation at the Biocuration Conference 2025
The Biocuration 2025 conference was held at the Stowers Research Institute in Kansas City, Missouri, chaired by Sofia Robb, a valued Monarch collaborator. The conference brought together over 100 attendees, both in-person and virtual, for engaging discussions on the current state of biocuration, developments using AI, and the future of biocuration.
The Monarch Initiative was well represented at the conference, with several team members leading workshops and delivering plenary talks highlighting the Initiative’s current efforts.
Collaborative Biocuration Workflows: Lessons learned from the OBO Academy
Facilitated by Sabrina Toro and Nicole Vasilevsky, the Collaborative Biocuration workshop brought together nearly 30 participants to explore best practices in open, community-driven biocuration. The session emphasized lessons from the Open Biomedical and Biological Ontology (OBO) Foundry, spotlighting the OBO Academy’s success in providing accessible, version-controlled, and high-quality training resources. Examples of successful collaborative biocuration projects included PomBase’s author-driven curation model, the Gene Ontology Consortium’s coordinated collaborative curation, and WormBase’s ACKnowledge project, which leverages AI to streamline author contributions. Participants discussed the importance of transparency, openness, low barriers to participation, data quality control, and education as key to the success of collaborative projects. They also explored the need for a centralized, open-access resource hub for basic biocuration training, akin to the OBO Academy. Key next steps include improving the accessibility of training material, developing certification pathways, and fostering regular community calls to share tools and knowledge across the field. Slides and materials are available at tislab.org/collaborativebiocuration and oboacademy.github.io/obook.
uPheno workshop: Enhancing the Unified Phenotype Ontology to Support Cross-Species Phenotype Interoperability
Led by Sue Bello, Nico Matentzoglu, and Arwa Ibrahim, the 4th uPheno workshop engaged participants from various ontologies and databases, including the plant community, the Rat Genome Database (RGD), the Mammalian Phenotype Ontology (MP), the Zebrafish Information Network (ZFIN), and the Database Center for Life Science (DBCLS), among others. The workshop focused on two main topics: 1) Improving the upper-level structure of the ontology and 2) Identifying integration and organization issues. Discussions also explored the current challenges in browsing the ontology, uPheno’s classification philosophy, reference ontology dependency, the utility of patternization for complex phenotypes, improving term labels and human-centric definitions, as well as leveraging AI to identify classification discrepancies. Future plans include working with plant ontology curators to integrate plant terms and developing collaborations with the RGD. Participants were encouraged to join the uPheno slack community, attend uPheno monthly calls, and create tickets in uPheno’s GitHub repository. Slides are available here.
Monarch Initiative: A Collaborative Journey
Sabrina Toro, Lead Monarch Curator, was invited to present at the “Maximizing Community Curation for the Benefit of All” workshop, hosted by WormBase researchers Kimberly Van Auken and Daniela Raciti, along with several other community curation projects. The presentation focused on the involvement of a global network of experts and community members, with a wide range of knowledge and skills, in all Monarch projects. The keys to success discussed were largely consistent with those discussed during the “Collaborative Biocuration Workflow” workshop. A particular focus was placed on the OBO Academy, along with strategies to simplify community contributions and foster a broader community (e.g. via web-based ways to propose ontology changes, language translation, and layperson terms).
Harnessing Community Power for Long-Term Success of the Mondo Disease Ontology
As part of the main conference, in the “Imageomics and Data Standards & Ontologies” session, Sabrina Toro presented Monarch’s strategies for harnessing community effort to keep Mondo up-to-date and how the community remains an active participant in the ongoing development of Mondo. Sabrina reported on the automated, quality-controlled, and scalable process that incorporates disease information from community collaborators directly into Mondo, using the maintenance of the rare disease subset as an example. The importance of recording information provenance and attribution was emphasized. Audience members were especially interested in the success of this community participation, noting that such community involvement is crucial for the future of biocuration projects.
Building the Lighthouse: Guiding LLM-Powered Biocuration with Domain Knowledge and Context
In Day 1’s session on Artificial Intelligence, Harry Caufield presented strategies and tools developed within Monarch to work with Large Language Models (LLMs) on biocuration efforts. LLMs can be used to extract structured data from unstructured information through tools like OntoGPT, CurateGPT, and Aurelian. The latter two tools can assist human curators with curation through task-specific agents. Human-curated resources remain crucial in these tasks as they serve as “lighthouses” guiding the use of LLMs due to the issue of LLM “hallucinations.”
