Explore protein conformational space with variational autoencoder

File explanation

Main

1. align_trajs.py: align trajectories
2. maps_featurize.py: turn traj data into xyz features
3. training.py: train VAE/AE models
4. models.py: model architecture for encoder, decoder and VAE

Utils

1. PDB_process.py: extract PDB template and write new coordinates
2. DOPE.py: calculate DOPE score of the given PDB
3. evaluation.py: evaluate the performance of VAE/AE models

Dependencies

The dependencies are listed in requirements.txt file.

For DOPE score calculation, modeller is required and can be downloaded here.

License

Apache-2.0