This repo is dedicated to the development of the CryoLike Python library, a fast and accurate algorithm for computing the likelihood of cryo-EM images given a cryo-EM map and atomic model. The algorithm is described in this bioRxiv paper https://www.biorxiv.org/content/10.1101/2024.10.18.619077v1
The Python library requires the following dependencies:
- numpy
- scipy
- matplotlib ## for plotting
- pytorch ## for GPU acceleration
- finufft ## for non-uniform FFT (https://finufft.readthedocs.io/en/latest/)
- cufinufft ## GPU version of finufft
- tqdm ## for progress bar
- mrcfile ## for reading MRC files
- mdtraj ## for reading PDB files and atom selection
If you are using virtual environment,
python3 -m venv ./venv/
source ./venv/bin/activate
pip install numpy scipy matplotlib tqdm mrcfile mdtraj
pip3 install torch torchvision
pip install finufft
pip install cufinufft
Note that the torch installation might depend on your system, please refer to the pytorch official website for the installation guide https://pytorch.org/get-started/locally/.
To install the Python library, clone the repository and run the following command in the root directory of the repository:
pip install .
To test the installation, run the following commands:
pip install pytest
pytest
For usage of the CryoLike algorithm, please refer to the example directory.
To run the examples, run the following command:
python3 make_templates_from_inputs_example.py
python3 convert_particle_stacks_example.py
python3 run_likelihood_example.py
which makes templates from the input files, converts particle stacks to templates, and runs the likelihood calculation, respectively. If the examples run successfully, you can find the output files in the output directory.
To plot the results, run the following command:
python3 plot_example.py
then you can find the plots in the output/likelihood directory.
Data used in the examples and test cases are retrieved from the following sources:
- EMPIAR-10026
- EMD: 2788
- PDB ID: 1UAO
- PDB ID: 4V1W