DeepMASC

A deep learning based tool to automatically select the best reconstructed 3D maps within a group of maps.

Features

• AutoSelect3D: Automatically select the best reconstructed 3D maps from RELION Class3D/InitialModel outputs
• AutoContour: Automatically determine contour levels and generate masks for cryo-EM maps
• Mask Evaluation: Evaluate 3D mask quality based on FSC criteria from RELION PostProcess jobs
• Module System: Optional environment module system for easy command-line access without manual activation

Example Data

The repository includes example data in the examples/ folder:

• examples/InitialModel/job015/: InitialModel job output with initial_model.mrc
• examples/Class3D/job016/: Class3D job output with 4 classes (run_it025_class001.mrc through run_it025_class004.mrc) and particle data (run_it025_data.star)

These examples can be used to test the functionality of DeepMASC without requiring your own data.

Credit: Example data is taken from the RELION Single Particle Analysis Tutorial.

Installation

Details

Clone the repository:

git clone https://github.com/AntiMatter568/DeepMASC

Install conda/mamba:

If you don't have conda or mamba installed, choose one of the following options:

Option 1: Miniconda (Recommended)

# Download and install Miniconda
wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh
# Follow the prompts and restart your terminal

Option 2: Miniforge (Community-driven, includes conda-forge by default + mamba)

# Download and install Miniforge (includes both conda and mamba)
wget https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-Linux-x86_64.sh
bash Miniforge3-Linux-x86_64.sh
# Follow the prompts and restart your terminal

For other operating systems, visit:

• Miniconda Downloads
• Miniforge Releases

Create conda environment:

cd <Your Installation Folder> # default is DeepMASC
conda env create -f environment.yml

# If using miniforge with mamba (faster alternative):
mamba env create -f environment.yml

Activate the environment (optional):

# Optional: You can activate the environment if you prefer
# conda activate DeepMASC

# Note: Environment activation is the same for both conda and mamba
# Alternatively, you can use the direct Python path as shown in examples

Configure config.py:

After setting up the conda environment, you need to configure the config.py file to point to your conda environment's Python executable:

1. Find your conda environment's Python path:

   # Method 1: Using conda info
   conda info --envs
   
   # Method 2: Check Python path after activating environment
   # conda activate DeepMASC
   # which python

2. Edit config.py file: Open config.py in the DeepMASC directory and update the CONDA_PYTHON_PATH variable:

   # Update this line with your actual conda environment path
   CONDA_PYTHON_PATH = "/path/to/your/conda/envs/DeepMASC/bin/python"

Install UCSF Chimera (Required):

UCSF Chimera is required for map resampling functionality in AutoContour and map visualization.

Installation Options:

1. Download and install from official website:

   • Visit: https://www.cgl.ucsf.edu/chimera/download.html
   • Download the appropriate version for your operating system
   • Follow the installation instructions for your platform

2. For Linux users:

   • Download the Linux version from the official website
   • Make the downloaded file executable and run the installer:

   # After downloading the .bin file from the website
   chmod +x chimera-*.bin
   ./chimera-*.bin

3. Verify installation:

   # Test that chimera command is available
   chimera --version
   
   # Test nogui mode (used by DeepMASC scripts)
   chimera --nogui --help

Note:

• Chimera is required for AutoContour resampling functionality
• Make sure the chimera command is available in your system PATH
• DeepMASC will fail if Chimera is not properly installed when using AutoContour features

Documentation

• AutoSelect3D Documentation - Automatically select the best 3D maps from Class3D/InitialModel outputs
• AutoContour Documentation - Automatically determine contour levels and generate masks
• Mask Evaluation Documentation - Evaluate 3D mask quality based on FSC criteria
• Module System Guide - Quick start guide for using DeepMASC with environment modules
• Module Installation Guide - Detailed installation instructions for the module system