Capra

Analysis toolchain for IBEX multiplex fluorescence microscopy images.

What it does

• Selectivity filtering to suppress autofluorescence and isolate true marker signal in multiplex IBEX data
• Interactive viewer (PyQt5) for real-time parameter tuning across all channels
• Cell segmentation via StarDist on DAPI nuclear channel

Setup

pip install -r requirements.txt

Place your registered IBEX .tif stacks in the project directory.

Quick start

python scripts/selectivity_viewer.py

Key concept: selectivity filtering

Real cell signal is bright in one specific channel. Autofluorescence and vessel bleed-through are bright across many channels. The selectivity metric exploits this:

selectivity = target / (mean_of_other_channels + epsilon)

High selectivity = real signal. Low selectivity = junk. The filtered image is target * normalized_selectivity.