Overview This python package contains a complete workflow and data managment suite to bring raw images of brain sections all the way through network analysis using ABBA/QuPath to align to a reference atlas.
Stages
- Preprocessing - prepare raw image for ABBA
- Configure data organization scheme
- Extract downsampled images apply order, rotate, flip, proofread, and rename images to unified format
- Extract/convert raw fullsize images to OME-TIFF + apply formatting
- Alignment
- create Qupath projects for each animal, align in ABBA, export back to QuPath, then run QuPath scripts to export registrations to open format
- Nuclei detection
- Nuclei quantification using StarDist, perform colocalization across image channels, and localize nuclei to atlas coordinates
- Data compilation
- Threshold nuclei based on morphological properties, reduce nuclei detections to counts per region/hemisphere
- Analysis
- Graph nuclei species per region across conditions (treatment, sex, etc.)
- PCA analysis to isolate differences between conditional groups
- Network analysis and interactive visualization - battery of graph analytics using NetworkX and Bokeh
- Volumetric renderings of aggregated nuclei counts