Towards a Modular Research Framework for Ontological Graphs

Ontologies are fundamental for structuring domain knowledge and ensuring semantic interoperability across scientific disciplines. However, many existing tools for ontology exploration and integration are either too rigid, overly complex for simple tasks, or exhibit a high memory footprint that limits scalability with large dataset. OntoGraph is a lightweight, open-source Python framework designed to overcome these limitations by offering a modular, graph-based architecture for efficient ontology exploration, querying, and integration. The framework follows a modular design, separating responsibilities between components and aiming for loose integration. This allows researchers to plug in new ontology formats and navigate between them without touching the core logic. Performance is a key design goal: OntoGraph currently integrates efficient graph operations, enabling scalable analysis of large ontologies. By combining modularity with high-performance computation, OntoGraph promotes sustainable, maintainable, and reproducible software, fully aligned with research software engineering (RSE) best practices. OntoGraph is a collaboration between Saezlab and the Scientific Software Center. Its first its first application is in algorithms processing metabolomics prior knowledge and features. From the start, the project emphasizes RSE best practices: automated testing, open-source development on GitHub, and clear documentation. Even at this early stage, the framework demonstrates potential as a reusable foundation for ontology-driven workflows, supporting integration and analysis across different research domains. At deRSE26, we will present the design rationale, the initial implementation, and the vision for this project. The poster will illustrate how modular software architecture could accelerate the development of research infrastructure tools while fostering interdisciplinary collaboration.