Nanoparticles used as drug carriers diffuse at different rates depending on their interactions with the surrounding medium. In the colon, this medium is colonic mucus, a complex biological barrier that strongly influences particle transport.
In this project, we investigated how drug-carrying particles move within both native and artificial mucus. Our objective was to quantify particle diffusivity by calculating the mean squared displacement (MSD) from particle trajectories. Because MSD analysis is highly dependent on accurate trajectory data, improving particle tracking was a critical component of the study.
To address this, we developed an advanced image-analysis pipeline capable of isolating and tracking nanoparticles in the heterogeneous background of mucus. The pipeline generated high-quality quantitative trajectory data, which then served as input for downstream mathematical modeling and machine-learning analysis.
Read the full publication: Machine learning framework for investigating nano- and micro-scale particle diffusion in colonic mucus., Tjakra, M., Lidayová, K., Avenel, C. et al., J Nanobiotechnol 23, 583 (2025).
The pre-processing of the data is done in FIJI software. To install FIJI follow the instructions on the FIJI website.
The tracking algorithm is implemented in Python using Trackpy toolkit.
Install the conda package, dependency and environment manager.
Then create the ChristelBergstrom2024-1 conda environment:
cd <path to your 'ChristelBergstrom2024-1' directory>
conda env create -f environment.yml
This will install all necessary project dependencies.
See trackpy for further information about the Trackpy toolkit.
Use FIJI macro for performing the batch pre-processing of all the data. The pre-processing consists of morphological top-hat operation with circular structuring element of size 10 and conversion into an 8-bit image.
Copy all project data to the data directory (or use symbolic links).
Then run Jupyter Lab from within the ChristelBergstrom2024-1 conda environment:
cd <path to your 'ChristelBergstrom2024-1' directory>
conda activate ChristelBergstrom2024-1
jupyter-lab
All analysis notebooks can be found in the notebooks directory.
See Initial tests notebook for detailed explanation of how to tune the tracking parameters on one data example.
Use Tracking notebook for performing the batch processing on all the data. Copy the parameter settings tuned in the previous notebook and update the input and output folder location.
If you find a bug, please raise an issue.
SciLifeLab BioImage Informatics Facility (BIIF)
Developed by Kristína Lidayová and Christophe Avenel