2D Confocal Microscopy Cell Image Segmentation and Quantitative Analysis

This repository contains a Python script for processing multi-channel microscopy images. It uses libraries such as cellpose, skimage, tifffile, and numpy to perform instance segmentation and quantitative analysis of cytoplasm, nuclei, mitochondria, and 2SC staining. The script generates processed images, segmentation overlays, and a CSV file with intensity and colocalisation data.

Features

• Image Preprocessing: Loads .tiff microscopy images with 4 channels, scales pixel intensities, and normalises each channel.
• Instance Segmentation:
  • Uses Cellpose for segmenting cytoplasm and nuclei.
  • Segments mitochondria via percentile-based intensity thresholding.
• Quantitative Analysis:
  • Extracts 2SC intensity from cytoplasm and nucleus.
  • Calculates colocalisation between mitochondria and 2SC staining within the cytoplasm.
  • Computes cell-wise areas for cytoplasm, nucleus, and mitochondria.
• Overlay Generation: Creates and saves overlay images of segmented regions.
• Batch Processing: Handles multiple .tiff images, processing each in sequence and saving the results in structured folders.

Dependencies

To run this script, install the following Python libraries:

pip install numpy tifffile scikit-image matplotlib opencv-python cellpose

Usage

Step 1: Prepare Image Data

Place your .tiff microscopy images in a directory (e.g., /path/to/images/). Ensure the images contain four channels (e.g., cytoplasm, nucleus, mitochondria, and 2SC staining).

Step 2: Modify the Script

Before running the script, modify the following parameters to correspond to the correct image channels:

cyto_channel: Cytoplasm channel (e.g., 1) nuclei_channel: Nucleus channel (e.g., 0) mito_channel: Mitochondria channel (e.g., 2) sc_channel: 2SC stain channel (e.g., 3)

Step 3: Run the Script

python process_images.py

Example usage within the script:

if __name__ == '__main__':
    
    img_path = '/path/to/images/'  # Directory with input images
    output_path = '/path/to/output/'  # Directory to save output

    nuclei_channel = 0  # Channel for nucleus
    cyto_channel = 1  # Channel for cytoplasm  
    mito_channel = 2  # Channel for mitochondria
    sc_channel = 3  # Channel for 2SC staining
    
    # Process the images and output results
    process_images(img_path, output_path, cyto_channel, nuclei_channel, mito_channel, sc_channel)

Step 4: Output

The script generates the following outputs:

Processed .tiff images with segmentation masks for cytoplasm, nucleus, and mitochondria.
Overlay images combining segmented regions for visualisation.
A CSV file (Intensity_Colocalisation_data.csv) containing quantitative data for intensity and colocalisation analysis.

Example Output

Sample output files include:

Processed Images: *_Processed_Image.tiff
Segmentation Masks: *_Cytoplasm_Mask.tiff, *_Nucleus_Mask.tiff, *_Mitochondria_Mask.tiff
Overlay Images: *_Segmentation_Overlay.tiff, *_MitoCyto2SC_Overlay.tiff
CSV File: Intensity_Colocalisation_data.csv

Example CSV Output

The CSV file contains the following columns:

Cytoplasm: Mean 2SC intensity in the cytoplasm.
Nucleus: Mean 2SC intensity in the nucleus.
Mito-2SC Colocalisation: Fraction of mitochondria colocalising with 2SC.
2SC-Mito Colocalisation: Fraction of 2SC colocalising with mitochondria.

Key Functions

load_image(image_path): Loads and normalises 4-channel .tiff images.
segment_instance(image, model_type, diameter, channels): Segments cytoplasm or nucleus using Cellpose.
segment_mitochondria(image, mito_channel_idx, percentile): Segments mitochondria based on intensity thresholding.
extract_cytoplasmic_2sc_intensity(image, cytoplasm_mask, nucleus_mask, mito_mask, sc_channel_idx): Extracts cytoplasmic 2SC intensity, excluding the nucleus and mitochondria.
calculate_colocalisation_in_cytoplasm(mito_mask, sc_mask, cyto_mask): Calculates colocalisation between mitochondria and 2SC staining.
overlay_segmentation(image, cytoplasm_mask, nucleus_mask): Generates overlay images of segmented regions.