Optimal Transport Based Graph Kernels for Drug Property Prediction (OTGK_DP)

This is the complete code for the OTGK_DP framework to predict ADMET durg properties. The work is pubished in the following papers:
(Paper)

@INPROCEEDINGS{10504311,
  author={Aburidi, Mohammed and Marcia, Roummel},
  booktitle={2024 IEEE First International Conference on Artificial Intelligence for Medicine, Health and Care (AIMHC)}, 
  title={Wasserstein Distance-Based Graph Kernel for Enhancing Drug Safety and Efficacy Prediction *}, 
  year={2024},
  volume={},
  number={},
  pages={113-119},
  keywords={Drugs;Proteins;Adaptation models;Toxicology;Pharmacodynamics;Predictive models;Safety;Optimal Transport;Wasserstein Distance;Graph Matching;Drug Discovery;ADMET Properties},
  doi={10.1109/AIMHC59811.2024.00029}}

Comments/Bugs/Problems: [email protected] [email protected]

December, 2023. Initial release

Find the Colab tutorial in the tutorial folder to run this repository and to predict the drug properties.
The data used in this project is Therapeutics Data Commons - TDC)
To list all of the datasets in AMDE, run the following

from tdc import utils
utils.retrieve_dataset_names('ADME')

To list all of the datasets in TOX, run the following

from tdc import utils
utils.retrieve_dataset_names('TOX')

Make sure to install the following dependincies:

PyTDC
rdkit
torch_geometric
POT
py3Dmol

Main Results:

Cite as:

@inproceedings{aburidi2024_1,
   author = {M. Aburidi and R. Marica},
   journal = {Scientific Reports},
   title = {Optimal Transport-Based Graph Kernels for Drug Property Prediction},
   url = {},
   year = {2024},
}

@INPROCEEDINGS{10504311,
  author={Aburidi, Mohammed and Marcia, Roummel},
  booktitle={2024 IEEE First International Conference on Artificial Intelligence for Medicine, Health and Care (AIMHC)}, 
  title={Wasserstein Distance-Based Graph Kernel for Enhancing Drug Safety and Efficacy Prediction}, 
  year={2024},
  volume={},
  number={},
  pages={113-119},
  keywords={Drugs;Proteins;Adaptation models;Toxicology;Pharmacodynamics;Predictive models;Safety;Optimal Transport;Wasserstein Distance;Graph Matching;Drug Discovery;ADMET Properties},
  doi={10.1109/AIMHC59811.2024.00029}}