OntoPlot: A Novel Visualisation for Non-hierarchical Associations in Large Ontologies

Ying Yang, Michael Wybrow, Yuan-Fang Li, Tobias Czauderna, Yongqun He

Abstract

Ontologies are formal representations of concepts and complex relationships among them. They have been widely used to capture comprehensive domain knowledge in areas such as biology and medicine, where large and complex ontologies can contain hundreds of thousands of concepts. Especially due to the large size of ontologies, visualisation is useful for authoring, exploring and understanding their underlying data. Existing ontology visualisation tools generally focus on the hierarchical structure, giving much less emphasis to non-hierarchical associations. In this paper we present OntoPlot, a novel visualisation specifically designed to facilitate the exploration of all concept associations whilst still showing an ontology's large hierarchical structure. This hybrid visualisation combines icicle plots, visual compression techniques and interactivity, improving space-efficiency and reducing visual structural complexity. We conducted a user study with domain experts to evaluate the usability of OntoPlot, comparing it with the de facto ontology editor Protégé. The results confirm that OntoPlot attains our design goals for association-related tasks and is strongly favoured by domain experts.

Introduction

Ontologies serve as the backbone of modern biomedical research, organizing vast amounts of knowledge into formal frameworks. However, as these datasets grow to include hundreds of thousands of classes, traditional visualization methods struggle to keep pace. Most existing tools focus almost exclusively on the subsumption hierarchy (the "is-a" relationship). This creates a significant gap in visualizing non-hierarchical associations, such as drug-side effect links. In standard editors like Protégé, these relationships are often buried in text-based side panels, making it nearly impossible for researchers to identify "class effects"—patterns where an entire group of concepts shares a specific association.

The OntoPlot Approach

To address these limitations, the authors introduce OntoPlot, a hybrid visualization system that integrates the top-down structure of an icicle plot with advanced visual compression. The system is designed to maintain the hierarchical context while prioritizing the visibility of complex associations.

Methods and Visual Compression

OntoPlot employs several novel techniques to manage visual complexity:

  • Hybrid Icicle Plot: Instead of traditional boxes, OntoPlot uses circle glyphs. To optimize horizontal space, leaf nodes are "wrapped" into grids rather than being stretched in a single line.
  • Automated Visual Compression: This core innovation identifies "uninteresting" subtrees—those containing no relevant associations—and collapses them into distinct glyphs. This keeps the hierarchy visible but minimized, reducing cognitive load.
  • Interactive Labeling: To prevent "label soup," the tool uses greedy placement and allows users to manually drag or "pin" labels to maintain a clear mental map.
Visual compression techniques in OntoPlot: (a) class node, (b) thin block for chains, (c) square for leaf groups, and (d) interactive labels.

Key Insights: Visualizing the "Class Effect"

A major advantage of OntoPlot is its ability to visualize how associations "bubble up" through the hierarchy. By using a heatmap-like coloring system for the circle glyphs, domain experts can instantaneously distinguish between outliers and broad class effects. If an entire block of circles is colored, the association applies to the whole class; if only one is colored, it is an outlier. This semantic zooming allows the hierarchy to function as a navigable map rather than a barrier to discovery.

The OntoPlot interface visualizing the OCVDAE dataset, showing hierarchical structure, compressed subtrees (triangles/squares), and association heatmaps.

Evaluation and Results

The researchers evaluated OntoPlot using the Ontology of Cardiovascular Drug Adverse Events (OCVDAE), a dataset containing 4,589 classes and over 20,000 non-hierarchical associations. A user study with 12 domain experts compared OntoPlot to the industry-standard Protégé.

  • Accuracy: OntoPlot significantly outperformed Protégé in complex tasks, such as identifying parent classes with the highest density of specific associations.
  • Speed: For association-related tasks, OntoPlot was substantially faster, often completing in seconds what took minutes in Protégé.
  • Expert Preference: While users found Protégé slightly easier for basic navigation due to familiarity, they overwhelmingly preferred OntoPlot for any task involving associations, citing its compact view and intuitive interface.

Conclusion

OntoPlot demonstrates that hierarchy should be a backbone for context, not a cage for data. By intelligently compressing irrelevant branches and highlighting non-hierarchical links, it provides a powerful new tool for exploring large-scale knowledge bases. Future work will focus on adding a global "mini-map" and supporting classes that belong to multiple parent lineages simultaneously.

Summary updated

Abstract

Ontologies are formal representations of concepts and complex relationships among them. They have been widely used to capture comprehensive domain knowledge in areas such as biology and medicine, where large and complex ontologies can contain hundreds of thousands of concepts. Especially due to the large size of ontologies, visualisation is useful for authoring, exploring and understanding their underlying data. Existing ontology visualisation tools generally focus on the hierarchical structure, giving much less emphasis to non-hierarchical associations. In this paper we present OntoPlot, a novel visualisation specifically designed to facilitate the exploration of all concept associations whilst still showing an ontology's large hierarchical structure. This hybrid visualisation combines icicle plots, visual compression techniques and interactivity, improving space-efficiency and reducing visual structural complexity. We conducted a user study with domain experts to evaluate the usability of OntoPlot, comparing it with the de facto ontology editor Prot{\'e}g{\'e}. The results confirm that OntoPlot attains our design goals for association-related tasks and is strongly favoured by domain experts.
BibTeX
@article{yang2019ontoplot,
  author = {Ying Yang and Michael Wybrow and Yuan-Fang Li and Tobias Czauderna and Yongqun He},
  title = {{OntoPlot: A Novel Visualisation for Non-hierarchical Associations in Large Ontologies}},
  journal = {IEEE Transactions on Visualization and Computer Graphics},
  year = {2019},
  doi = {10.1109/tvcg.2019.2934557},
}
Figure