AMP-BMS is a multiscale neural network potential trained on QM/MM data with electrostatic embedding for the simulations of biomolecules in the condensed phase. This repo was used for our recent publication together with a dataset for multiscale neural networks. This model is based on the AMP architecture and was used in previous work 1, 2.
Example usage is shown in here.
# Install environment.
conda env create -f amp.yml
conda activate amp_bms
# Install repo after cloning from github
pip install -e .
If you use this code, please cite our papers:
Moritz Thürlemann
Felix Pultar
Igor Gordiy
The AMP-BMS code is published and distributed under the MIT License.