PhotoFiTT: Phototoxicity Fitness Time Trial

A Quantitative Framework for Assessing Phototoxicity in Live-Cell

General description of the workflow

PhotoFiTT was designed to quantitatively analyse the impact that fluorescence light excitation has in cell behaviour. PhotoFiTT focuses on three different measurements: (1) Identified pre-mitotic cells, (2) Cell size dynamics and (3) Cell activity. These are the steps to follow to replicate the analysis:

Deep learning based analysis

Follow these steps to detect cells and pre-mitotic rounding events in the data.

1. Cell Detection and Quantification (deep learning-based image analysis: This processing is only applied to the first time point of each video.
   • Virtual Staining: Use ZeroCostDL4Mic/DL4MicEverywhere Pix2Pix notebook to train a virtual staining model that infers cell nuclei. Analyse the first frame of each video.
   • Nuclei Segmentation: Use ZeroCostDL4Mic/DL4MicEverywhere 2D StarDist notebook to apply the pretrained StarDist-versatile model to segment individual nuclei in the virtually stained images.
   • Initial Cell Quantification: Count the number of detected nuclei (Use notebook XXXXX.ipnynb to generate a CSV file with the counts). The number of detected nuclei serves as the baseline cell count for each field of view, enabling tracking of population dynamics over time.
2. Pre-mitotic Cell Identification (deep learning-based image analysis):
   • For CHO cells imaged with brightfield, you can use our trained StarDist model. Otherwise, manually annotate a representative image set and train a new StarDist model using the corresponding ZeroCostDL4Mic/DL4MicEverywhere notebooks.

Image data Analysis

1. Cell Size Analysis and Classification XXXXX.ipnynb
2. Quantification of Cellular Activity XXXXX.ipnynb

Data structure

1. The masks and the raw input, should be equally organised by folders, each folder for each condition to be analysed in a hierarchical manner. For example:

    -Raw-images (folder)
    |
    |--Biological-replica-date-1 (folder) [Subcaegory-00]
        |
        |--Cell density / UV Ligth / WL 475 light [Subcategory-01] 
           |
           |-- control-condition (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
           |
           |-- condition1 (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
           |
           |-- condition2 (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
        |
        |--Cell density / UV Ligth / WL 475 light [Subcategory-01]
        ...
    -Masks (folder)
    |
    |--Biological-replica-date-1 (folder) [Subcaegory-00]
        |
        |--Cell density / UV Ligth / WL 475 light [Subcategory-01] 
           |
           |-- control-condition (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
           |
           |-- condition1 (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
           |
           |-- condition2 (folder) [Subcategory-02] 
           |    |  file1.tif
           |    |  file2.tif
           |    |  ...
        |
        |--Cell density / UV Ligth / WL 475 light [Subcategory-01]
        ...
   

Package installation

• The code provides an environment.yaml file to create a conda environment with all the dependencies needed. Place your terminal in the photofitt folder. Use either conda or mamba:

  git clone https://github.com/HenriquesLab/photofitt.git
  cd photofitt
  mamba env create -f environment.yml  
  mamba activate photofitt
  

• ONCE PUBLISHED You can now install the package using pip install or conda as follows:

  • pip install photofitt
    

    or

  • conda install photofitt
    

• Meanwhile:

  • git clone https://github.com/HenriquesLab/photofitt.git
    cd photofitt
    python setup.py
    

    or

  • git clone https://github.com/HenriquesLab/photofitt.git
    cd photofitt
    pip install .
    

    or

  • git clone https://github.com/HenriquesLab/photofitt.git
    cd photofitt
    conda build conda-recipe/meta.yaml
    

Common error messages

• Error messages with lxml. Most probably you need to update developers tools in your system. Before anything, run in Mac M1:

  ```
  xcode-select --install
  ```
  

• If you were in Linux, you can run

  • sudo apt-get update
    sudo apt-get install libxml2-dev libxslt-dev python-dev