YZS-model: A Predictive Model for Organic Drug Solubility Based on Graph Convolutional Networks and Transformer-Attention.

Configuration of the YSZ-model environment

python==3.6
torch==1.7.1
scikit-learn==0.24.2
scipy==1.5.4
torch-geometric==1.7.0
einops==0.4.1
networkx==2.5.1
rdkit-pypi

Other packages can be found from requirements.txt.

Datasets

You can download the training and testing dataset from:

All datasets in the raw folders.

Useage

• smile2topology.py: Convert .csv files to datasets.
• model.py: The whole YZS-model.
• opti.py: Using searching package to find the perfect parameters.
• train.py: Training the YZS-model.
• test.py:Tests and evaluates the YZS model.

About

authors:

• Chenxu Wang, Shihezi University (PRC, Xinjiang)
  • e-mail:[email protected]
• Haowei Ming, Peking University (PRC, Beijing)
• Jian He, Xinjiang University (PRC, Xinjiang)
• Yao Lu, Shihezi University (PRC, Xinjiang)
• Junhong Chen, {South China University of Technology, NetEase Inc.} (PRC, {Guangdong, Zhejiang})

Statement:

• Part of code come from:

https://github.com/ziduzidu/CSDTI
https://github.com/ltorres97/FS-CrossTR
https://github.com/waqarahmadm019/AquaPred

When using the above-mentioned open-source code, we have already indicated this in the documents.