FM-Track

FM-Track is a feature-based fiducial marker tracking software for applications in cell mechanics.

Research methods in mechanobiology that involve tracking the deformation of fiducial markers in the vicinity of a cell are increasing in popularity. Here we present a software called FM-Track, a software to facilitate feature-based particle tracking tailored to applications in cell mechanics. FM-Track contains functions for pre-processing images, running fiducial marker tracking, and simple post-processing and visualization. We expect that FM-Track will help researchers in mechanobiology and related fields by providing straightforward and extensible software written in the Python language.

For a lengthier description, please refer to the technical overview.

Installation

To install FM-Track, it is recommended to use Pip within a conda environment. Anaconda allows you to create environments to make software dependencies and installations easier to manage. By installing FM-Track into a conda environment, Anaconda automatically configures the proper dependencies for you and installs them all into the correct location. Once installed, FM-Track can be run using Python import statements. This allows you to import FM-Track functions from the Anaconda installation folder so they can be used in an external file.

Currently, FM-Track is only compatible with Python <= 3.6. To create a compatible environment, use the following commands:

conda create -n fmtrack python=3.6 anaconda
conda activate fmtrack
conda install -c conda-forge latexcodec
conda install pip

To install FM-Track, download the zip file for the repository, unzip the folder, and navigate into the folder inside of a bash shell (such as terminal) using the following command:

cd <path-to-folder>

To complete the installation, run the following command:

sudo python setup.py install

Once running this command, you may delete the downloaded folder if you wish, however we strongly recommend keeping the examples folder in an easily accessible location. The source code folder (located in FM-Track-master/fmtrack) has now been copied into the proper Anaconda environment location. In other words, changing the source code will only affect the execution of scripts implementing FM-Track if you change the code now located in the Anaconda environment folder.

Testing and Usage

The easiest way to get started with FM-Track is by running and modifying the test.py script. This script runs FM-Track on a set of sample data so you can see how it works on test data before trying it on your own. To properly use this file, copy the data folder to your Desktop (or somewhere easily accessible), then change the root_directory variable at the top of test.py to store the path of this data folder. Next, run test.py using an IDE or in terminal by calling the following commands:

conda activate fmtrack
python <path-to-test.py>

When importing functions from FM-Track, test.py looks for the files installed into your anaconda environment folder. Because of this, you must activate the environment every time you want to call FM-Track functions. To do this, use the command conda activate fmtrack.

FM-Track uses the input_info class to store all your filepaths. The test.py file explains this in greater detail. Additionally, input_info stores the values of other tunable parameters. The test.py file explains this as well. The parameters that can be modified are as follows:

• Color channels of beads and cell
• Dimensions of FOV
• Cellular threshold for analyzing boundaries of cellular material
• Number of feature vectors used
• Number of nearest beads analyzed
• Tracking type (translation correction or not)
• Cell buffers (for both the regular and translation correcting steps)
• Data plotting options
• Terminal output options

For a greater description of all of these, please refer to the technical overview.

Built With

• py-earth - Used to compute multivariate adaptive regression splines
• scikit-learn - Used to generate Gaussian Process Regression models
• pyvista - Used to generate 3D deformation field graphs

Authors

• Emma Lejeune
• Alex Khang
• Jake Sansom

License

This project is licensed under the MIT License - see the LICENSE file for details