HPC-Organoid-Quant

Organoid Detection and GFP Quantification

This project provides a complete and modular pipeline to detect organoids from brightfield microscopy images, with optional quantification of GFP-positive regions from corresponding fluorescence images.

Features

• Works with brightfield-only images or brightfield + GFP
• Detects irregular and fragmented organoids as a single object
• Computes key morphological features:
  • Area
  • Centroid
  • Perimeter
  • Circularity
• If GFP is provided:
  • Quantifies GFP+ area
  • Calculates mean and total GFP intensity
• Generates annotated overlay images for visualization

More info on the features: https://github.com/arka2696/HPC-Organoid-Quant/blob/main/FeatureExplanation.md

Project Structure

Input folder structure:

Organised_files/
└── B16 Spheroid_10% FBS/
    └── B16_control/
        └── B16 day 3/
            ├── BRIGHTFIELD_16173.jpg
            ├── BRIGHTFIELD_16174.jpg
            └── ...

Output folder structure:

OUTPUT/
├── B16 Spheroid_10% FBS/
│   └── B16_control/
│       └── B16 day 3/
│           ├── masks/
│           │   ├── mask_16173.png
│           │   └── ...
│           └── overlays/
│               ├── overlay_16173.png
│               └── ...
│
└── organoid_gfp_analysis.csv   ← all image results combined here

---

Getting Started

1. Clone the Repository

git clone https://github.com/yourusername/Organoid_GFP_Analysis.git
cd Organoid_GFP_Analysis

2. Launch in Google Colab

Best used with Google Drive integration.

• Open notebooks/organoid_analysis.ipynb
• Update paths to your own brightfield and GFP images
• Run all cells

3. Launch with Conda or Python venv

• Create the conda env with this command:

  conda env create -f environment.yml
  conda activate organoid-env

• OR, create a python venv with thsi command:

  python -m venv organoid-env
  source organoid-env/bin/activate  # or .\organoid-env\Scripts\activate on Windows
  pip install -r requirements.txt

4. Creating a loadable jupyter-kernel: This is needed as if we are runing the code from jupyter lab or VS-code we need to load python kenrnel

• A kernel can be created by the following command:

  python -m ipykernel insatll --user --name organoid-env --display-name "organoid-env"

5. Launching the organoid-env environment

• We can launch the organoid-env conda env in two ways: 1. via anaconda prompt or 2. we can directly load the conda kernel in VS-Code

• 1. For anaconda prompt first open anaconda prompt:
• 2. In the command line/terminal type in the following command:

  conda activate organoid-env

  then

  jupyter lab

  The organoid-env is loaded by default, but if its not loaded, we can select the kernel from the dropdown box on the top right hand.

• 3. Now you can directly run the .ipynb notebooks within the jupyter-lab or directly execute the python script from the command line.

Output Example

• .csv file with organoid measurements

• Overlay image:

  • 🔵 Organoid outline
  • 🔴 GFP+ region (if applicable)

---

Dependencies

All dependencies are standard and available in Google Colab:

• OpenCV
• NumPy
• Matplotlib
• scikit-image
• Pandas

If you want to run locally, create a requirements.txt:

pip install -r requirements.txt

---

Notes

• The script accepts a broad format of images
• If your organoids are fragmented, the pipeline includes morphological merging to treat them as one object
• GFP quantification thresholds can be adjusted in the notebook

---

Contact

For questions or suggestions, feel free to open an issue or reach out via GitHub.

---

Citing the repo:

v1.0.0 :

[Latest] v1.1.0 :

Cite as: Arkajyoti sarkar. (2025). HPC-Organoid-Quant (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.15738206

License

This project is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License.