seokhokang / nmr_mpnn

Neural Message Passing for NMR Chemical Shift Prediction

nmr_mpnn

Tensorflow implementation of the model described in the paper Neural Message Passing for NMR Chemical Shift Prediction

Our PyTorch implementation is available at https://github.com/seokhokang/nmr_mpnn_pytorch

Components

• preprocessing.py - script for data preprocessing
• run_train.py - script for model training
• run_eval.py - script for model evaluation
• MPNN.py - model architecture
• util.py

Data

• NMRShiftDB2 - https://nmrshiftdb.nmr.uni-koeln.de/
• The datasets used in the paper can be downloaded from
  • https://jcheminf.biomedcentral.com/articles/10.1186/s13321-019-0374-3
  • https://static-content.springer.com/esm/art%3A10.1186%2Fs13321-019-0374-3/MediaObjects/13321_2019_374_MOESM2_ESM.gz

Dependencies

• Python
• TensorFlow
• RDKit
• NumPy
• scikit-learn
• sparse

Citation

@Article{Kwon2020,
  title={Neural message passing for NMR chemical shift prediction},
  author={Kwon, Youngchun and Lee, Dongseon and Choi, Youn-Suk and Kang, Myeonginn and Kang, Seokho},
  journal={Journal of Chemical Information and Modeling},
  volume={60},
  number={4},
  pages={2024-2030},
  year={2020},
  doi={10.1021/acs.jcim.0c00195}
}