STORM Analysis

The best way to run the analysis is on your local machine. On Mac:

• First time only:
  • Make sure python is installed by typing python --version or python3 --version in terminal.
  • Navigate to the folder you would like to run the analysis from using the cd command (e.g. cd Desktop)
  • Clone the repository: git clone https://github.com/cmackenzie29/storm.git
  • Install necessary packages: pip install -r requirements.txt
• To run the analysis:
  • Navigate to the storm folder (e.g. cd storm, cd Desktop/storm, etc.)
  • Run the program by typing python storm.py or python3 storm.py
    • The program will allow you to interactively select DBSCAN parameters that work for your data. Optionally, if you already know the DBSCAN parameters you would like to use, you can specify them as command-line arguments like python storm.py --eps=0.5 --min_samples=10
• If you do not know the colors present in your txt files, check by running python find_colors.py

Alternatively, you can run the analysis from Google Colab with your STORM files stored in Google Drive.

Analysis Modes

• Basic: Analyze all molecules in the image, without clustering
• Clusters with all molecules together: Analyze only the molecules in the image which are in clusters, where each cluster includes all specified colors
• Clusters with only one color: Analyze only the molecules in the image which are in clusters, where each cluster includes just one color

Outputs

• nndists (basic and clusters): Average distance from molecules of one color to each nearest molecule of another color, for every possible combination of colors
• synapses (clusters):
  • Average molecule count per synapse for each color
  • Average molecule density per synapse for each color
  • Average synapse volume for each color
  • Average molecule intensity per synapse for each color
• png files (clusters) showing the identified clusters for each image given the chosen DBSCAN parameters

Note: distance units are μm