Context Aware Template Matching (CATM) for Cryo-ET data analysis

CATM version 0.1

Introduction

This software is designed to assign models in crowded cryo-ET tomograms. Its primary function is to take pre-picked coordinates from other software as input and calculate the local cross-correlation coefficients (CCCs) with the provided templates. The software then integrates this information with geometric constraints between models to achieve highly accurate assignments. It also includes a clash resolver module to optimize the poses of models, with support for multi-template assignments.

Installation

Python version 3.5 or higher is required. It is recommended to create a separate environment for installation to avoid conflicts with other packages. The installation typically takes less than 5 minutes, depending on your network speed and hardware performance.

1. Creat a new vitual enviroment (optional)

conda create -n CATM python==3.10

2. Download the package from gitlab and change directory to the CATM folder

git clone https://git.biohpc.swmed.edu/s194231/CATM.git #(optional if you already have the files)
cd CATM

3. Install the package, if you want to develop base on this software, add -dev option

pip install .

4. After installation, activate the conda environment if not yet

conda activate CATM

Test the installation and Tutorial

Please refer to the Tutorial file for step-by-step instructions on testing the installation and using the software.

General Usage

• The program currently supports three modes: "CATM", "CATM-local", and "general TM".

1. Prepare your data, required files are:

• A tomogram file in .mrc format, typically produced by back-projection and low-pass filtered.
• A list of candidate particles from deep-learning-based pickers, manual picking, or other methods. Supported formats include .csv, .coord, and .star.
• A 3D CTF model from Warp or Relion (optional). If not provided, you must specify the missing wedge information.
• One or more template files in cubic volume format.
• Corresponding mask files for each template (optional). These must have the same dimensions as the templates. If not provided, a tight mask will be generated automatically (recommended).

2. Determine the contour level for template matching and clash resolver

• Run the counter_level_check.py file to determine the contour level for your template. This step only needs to be done once for similar datasets.

  python counter_level_check.py path_to_template.mrc [ctf_model or missing wedge info]

• Examine the contour level in Chimera/ChimeraX and choose the one with the least artifacts.

3.Copy the config.py file to your working directory and adjust the parameters

• The key parameters should be adjusted according to your needs. The program supports three modes: "CATM", "CATM-local", and "TM".

4. Run the program and analyze the results

CATM

Run the program in your working directory. The program will output results in .csv, .coord, and .star formats. A back-mapped tomogram will also be generated.

License

UTSW open source

Free Software for academia!