Cell ontology in an age of data-driven cell classification

Area of Science:

• Computational Biology
• Bioinformatics
• Cellular and Molecular Biology

Background:

• Large-scale data-driven cell classification projects, like the Human Cell Atlas, face challenges in making results searchable and accessible to biologists.
• Integrating massive datasets with existing knowledge of cell types, anatomy, and development is crucial for biological discovery.
• Cross-searching diverse single-cell analysis data requires standardized methods for data integration and retrieval.

Purpose of the Study:

• To demonstrate how formalizing cell ontologies can improve querying of data-driven cell classifications.
• To illustrate how ontologies can be extended by integrating outputs from data-driven cell classifications.
• To explore the role of structured annotation in making cell classification data searchable and queryable.

Main Methods:

• Utilized case studies from the mouse retina and Drosophila olfactory system.
• Applied formalization of cell ontologies to enhance data-driven cell classifications.
• Integrated outputs from data-driven cell classifications to extend existing ontologies.

Main Results:

• Demonstrated enhanced querying of data-driven cell classifications through formalized cell ontologies.
• Showcased the extension of ontologies by incorporating data-driven cell classification outputs.
• Provided worked examples from specific biological systems (mouse retina, Drosophila olfactory system).

Conclusions:

• Annotation with ontology terms is vital for making data-driven cell classifications searchable and queryable.
• Standardized formal patterns are required for structuring ontologies and annotations.
• Linking ontologies to data-driven classification outputs is essential for realizing their full potential.