Authors: Shengchao Liu*, Weitao Du*, Hannan Xu, Yanjing Li, Zhuoxinran Li, Vignesh Bhethanabotla, Divin Yan, Christian Borgs*, Anima Anandkumar*, Hongyu Guo*, Jennifer Chayes*

[Project Page] [ArXiv] [Datasets on HuggingFace] [Checkpoints on HuggingFace]

1. Environment

Conda

Setup the anaconda

wget https://repo.continuum.io/archive/Anaconda3-2019.10-Linux-x86_64.sh
bash Anaconda3-2019.10-Linux-x86_64.sh -b
export PATH=$PWD/anaconda3/bin:$PATH

Packages

Start with some basic packages.

conda create -n Geom3D python=3.9
conda activate Geom3D
conda install -y numpy networkx scikit-learn
conda install -y -c conda-forge rdkit
conda install -y pytorch==2.2 pytorch-cuda=12.1 -c pytorch -c nvidia
conda install -y -c pyg -c conda-forge pyg=2.5
conda install -y -c pyg pytorch-scatter
conda install -y -c pyg pytorch-sparse
conda install -y -c pyg pytorch-cluster

pip install ogb==1.2.1

pip install sympy

pip install ase

pip install lie_learn # for TFN and SE3-Trans

pip install packaging # for SEGNN
pip3 install e3nn # for SEGNN

pip install transformers # for smiles
pip install selfies # for selfies

pip install atom3d # for Atom3D
pip install cffi # for Atom3D
pip install biopython # for Atom3D

pip install cython # for pyximport 

conda install -y -c conda-forge py-xgboost-cpu # for XGB

pip install pymatgen  # for CIF loading
pip install h5py

pip install torch-ema

git clone git@github.com:chao1224/torchdiffeq.git
cd torchdiffeq

pip install MDAnalysis

pip install -e .

2. Datasets Preparation

We provide two ways to generate the datasets for MISATO.

1. We provide the script under data/MISATO to generate two sub-datasets, and you can check the data/README.md for more details.
2. You can download the datasets from zenodo and HuggingFace directly. 2.1. You can download the MISATO MD.hdf5 data from zenodo link, or use the following CMD:

wget -O data/MD/h5_files/MD.hdf5 https://zenodo.org/record/7711953/files/MD.hdf5

2.2. Then you can download the dataset from HuggingFace link provided by us.

The data folder structure looks like the following:

.
`-- MISATO_1000
|   `-- raw
|   |   `-- train_MD.txt
|   |   `-- test_MD.txt
|   |   `-- MD.hdf5
|   |   `-- val_MD.txt
`-- MISATO
|   `-- raw
|   |   `-- train_MD.txt
|   |   `-- test_MD.txt
|   |   `-- MD.hdf5
|   |   `-- val_MD.txt
`-- README.md
`-- MISATO_100
|   `-- raw
|   |   `-- train_MD.txt
|   |   `-- test_MD.txt
|   |   `-- MD.hdf5
|   |   `-- val_MD.txt

3. Scripts

Please check examples for semi-flexible binding experiments.

We have two types of tasks

• multi_traj
• single_traj and four ML methods
• VerletMD
• GNNMD
• DenoisinLD
• NeuralMD
  • --NeuralMD_binding_model=NeuralMD_Binding01 for NeuralMD ODE
  • --NeuralMD_binding_model=NeuralMD_Binding02 or --NeuralMD_binding_model=NeuralMD_Binding04 for NeuralMD SDE

4. Checkpoints

We provide the optimal checkpoints and corresponding hyperparameters at this HuggingFace link.

Cite Us

Feel free to cite this work if you find it useful to you!

@inproceedings{
    @article{liu2024NeuralMD,
    title={A Multi-Grained Symmetric Differential Equation Model for Learning Protein-Ligand Binding Dynamics},
    author={Liu, Shengchao* and Du, Weitao* and Xu, Hannan and Li, Yanjing and Li, Zhuoxinran and Bhethanabotla, Vignesh and Liang, Yan and Borgs, Christian* and Anandkumar, Anima* and Guo, Hongyu* and Chayes, Jennifer*},
    journal={arXiv preprint arXiv:2401.15122},
    year={2024}
}