Flexible and robust cell type annotation for highly multiplexed tissue images.

Area of Science:

• Computational biology
• Bioinformatics
• Digital pathology

Background:

• Accurate cell type identification is crucial for understanding tissue architecture.
• Existing methods for cell annotation often require extensive manual effort and reference datasets.
• Limitations in current tools hinder large-scale analysis of multiplexed imaging data.

Purpose of the Study:

• To develop an automated, unbiased, and accurate tool for cell type annotation in highly multiplexed images.
• To enable fine-grained cell classification across diverse antibody panels without retraining.
• To facilitate the analysis of spatial organization in human tissues.

Main Methods:

• Development of the Robust Image-Based Cell Annotator (RIBCA) tool.
• Application of RIBCA to annotate cell types in multiplexed imaging data.
• Utilizing a modular design for extensibility to new cell types and antibody panels.

Main Results:

• Successful annotation of over 3 million cells across more than 40 human tissue types.
• Demonstrated accuracy, automation, and unbiased performance of the RIBCA tool.
• Revealed intricate spatial organization patterns of various cell types within tissues.

Conclusions:

• RIBCA provides a robust solution for automated cell type annotation in complex imaging datasets.
• The tool's open-source nature and modular design promote wider adoption and further development.
• RIBCA significantly advances the study of tissue spatial organization and cellular heterogeneity.