Awesome
GraphDG
<img src="resources/confs.png" width="40%">A Generative Model for Molecular Distance Geometry<br> Gregor N. C. Simm, José Miguel Hernández-Lobato <br> Proceedings of the 37th International Conference on Machine Learning, Vienna, Austria, PMLR 119, 2020.<br> https://arxiv.org/abs/1909.11459
Installation
-
Create new Python 3.7 environment and activate:
virtualenv --python=python3.7 graphdg-venv source graphdg-venv/bin/activate
-
Install required packages and library itself:
pip install -r graphdg/requirements.txt pip install -e graphdg/
-
Install RDKit (2020.03.1).
Usage
-
Download and unpack ISO17 dataset
wget http://quantum-machine.org/datasets/iso17.tar.gz tar -xf iso17.tar.gz
-
Prepare dataset
python3 graphdg/scripts/parse.py --path=iso17
-
Train model and generate conformations
python3 graphdg/scripts/run.py --train_path=iso17_split-0_train.pkl --test_path=iso17_split-0_test.pkl
Reference
@inproceedings{Simm2020GraphDG,
title = {A Generative Model for Molecular Distance Geometry},
author = {Simm, Gregor N. C. and Hern\'andez-Lobato, Jos\'e Miguel},
booktitle = {Proceedings of the 37th International Conference on Machine Learning},
pages = {8949--8958},
year = {2020},
editor = {Hal Daumé III and Aarti Singh},
volume = {119},
series = {Proceedings of Machine Learning Research},
month = {13--18 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v119/simm20a/simm20a.pdf},
url = {http://proceedings.mlr.press/v119/simm20a.html}
}