ETH-ScopeM_Krupke

Fluorescent dye intra-tissue diffusion analysis
This code is published here: https://doi.org/10.1016/j.jconrel.2025.113947
Published version: "v1.0-paper"

Index

• Installation
• Usage
• Comments

Installation

Pease select your operating system

Windows

Step 1: Download this GitHub Repository

• Click on the green <> Code button and download ZIP
• Unzip the downloaded file to a desired location

Step 2: Install Miniforge (Minimal Conda installer)

• Download and install Miniforge for your operating system
• Run the downloaded .exe file
  • Select "Add Miniforge3 to PATH environment variable"

Step 3: Setup Conda

• Open the newly installed Miniforge Prompt
• Move to the downloaded GitHub repository
• Run one of the following command:

# TensorFlow with GPU support
mamba env create -f environment_tf_gpu.yml
# TensorFlow with no GPU support 
mamba env create -f environment_tf_nogpu.yml

• Activate Conda environment:

conda activate Krupke

Your prompt should now start with (Krupke) instead of (base)

MacOS

Step 1: Download this GitHub Repository

• Click on the green <> Code button and download ZIP
• Unzip the downloaded file to a desired location

Step 2: Install Miniforge (Minimal Conda installer)

• Download and install Miniforge for your operating system
• Open your terminal
• Move to the directory containing the Miniforge installer
• Run one of the following command:

# Intel-Series
bash Miniforge3-MacOSX-x86_64.sh
# M-Series
bash Miniforge3-MacOSX-arm64.sh

Step 3: Setup Conda

• Re-open your terminal
• Move to the downloaded GitHub repository
• Run one of the following command:

# TensorFlow with GPU support
mamba env create -f environment_tf_gpu.yml
# TensorFlow with no GPU support 
mamba env create -f environment_tf_nogpu.yml

• Activate Conda environment:

conda activate Krupke

Your prompt should now start with (Krupke) instead of (base)

Usage

process.py

Read .lif images from data_path folder, process and save outputs in a new folder named according to image name.

• Paths

- img_name         # str, image name ("all" for batch processing)
- data_path        # str, path to folder containing image(s) to process
- model_mass_path  # str, path to DL segmentation model

• Parameters

- df               # int, downscaling factor, should be kept at 30 (DL model)

• Outputs

- image.tif        # uint16, downscaled image
- prediction.tif   # float32, DL segmentation prediction
- mask.tif         # uint8, tissue segmentation mask
- outline.tif      # uint8, tissue surface mask
- metadata.txt     # df, original and downscaled pixel size in µm
- metadata.pkl     # df, original and downscaled pixel size in µm

correct.py

Read all processed images from data_path and display outline for manual corrections in Napari. Erase the undesired surface and press enter to save the corrected outline_hc.

• Paths

- data_path        # str, path to folder containing image(s) to process

• Parameters

- erase_size       # int, size in pixel(s) of erasing tool
- paint_size       # int, size in pixel(s) of painting tool

• Outputs

- outline_hc.tif   # uint8, corrected tissue surface mask

analyse.py

Read all processed images from data_path and compute Euclidean distance transform edt of outline_hc to measure fluorescence intensities according to the distance from the surface. If not 0, the baseline_pc paramater define the percentage of lowest values that will be considered to define a baseline to define a baseline which will be subtracted from the measured values.

• Paths

- data_path        # str, path to folder containing image(s) to process

• Parameters

- max_bin          # int, max bin distance in µm
- num_bins         # int, number of bins between 0 and max_bin
- baseline_pc      # float, percentage (0 to 100) of lowest values to be considered baseline

• Outputs

- image_bsub.tif   # float32, baseline subtracted image
- edt.tif          # float32, Euclidean distance transform of outline_hc
- display.tif      # RGB, image and outline_hc overlay 
- results.csv      # intensities acc. to distance in µm
- results.png      # plot of intensities acc. to distance in µm
- metadata.txt     # add baseline
- metadata.pkl     # add baseline

Comments