HERO: A hierarchy-aware analysis pipeline for reducing and refining whole-brain atlas-mapped cellular datasets

Area of Science:

• Neuroscience
• Bioinformatics
• Computational Biology

Background:

• High-throughput microscopy and machine learning enable whole-brain imaging.
• Analyzing large, hierarchical brain datasets presents significant challenges for biological interpretation.
• Inconsistent reporting methods hinder reproducibility in brain imaging studies.

Purpose of the Study:

• To develop a user-friendly workflow for reducing and organizing whole-brain cell detection data.
• To enhance the interpretability and accessibility of complex neuroimaging datasets.
• To standardize analysis methods for improved rigor and reproducibility.

Main Methods:

• Developed HERO (Hierarchy-aware Expression Region Organization), a data reduction workflow.
• Workflow performs hierarchy-aware selection, refinement, ranking, and visualization.
• Designed as a customizable, plug-in solution requiring minimal coding experience.

Main Results:

• HERO provides transparent, curated results from raw cell detection data.
• Enables efficient, unbiased region selection for streamlined statistical analysis.
• Successfully transforms large datasets into biologically meaningful insights.

Conclusions:

• HERO offers a standardized approach to reduce and analyze whole-brain cell detection datasets.
• Improves the effectiveness, interpretability, and accessibility of whole-brain imaging.
• Applicable to various atlas-mapped datasets with hierarchical information.