Awesome

Note:

This project contains several GNN-based models for protein-ligand binding affinity prediction, which are mainly taken from
PotentialNet: https://github.com/awslabs/dgl-lifesci/blob/master/python/dgllife/model/model_zoo/potentialnet.py
GNN_DTI: https://github.com/jaechanglim/GNN_DTI
IGN: https://github.com/zjujdj/InteractionGraphNet/tree/master
SchNet: https://pytorch-geometric.readthedocs.io/en/latest/_modules/torch_geometric/nn/models/schnet.html
EGNN: https://github.com/vgsatorras/egnn
RF-Score: The RF-Score features are extracted using OODT: https://github.com/oddt/oddt, we already provide the processed data in ./RF-Score/feature so that the project can be run directly.
Each baseline directory has a requirement.txt listing the version of the packages.

Dataset:

All data used in this paper are publicly available and can be accessed here:

• PDBbind v2016 and v2019: http://www.pdbbind.org.cn/download.php
• 2013 and 2016 core sets: http://www.pdbbind.org.cn/casf.php
  You can also download the processed data from https://zenodo.org/record/7490623#.Y60PTnZBxD8

Requirements:

matplotlib==3.3.4
networkx==2.5
numpy==1.19.2
pandas==1.1.5
pymol==0.1.0
rdkit==2022.9.2
scikit_learn==1.1.3
scipy==1.5.2
seaborn==0.11.2
torch==1.10.2
torch_geometric==2.0.3
tqdm==4.63.0

Usage:

We provide a demo to show how to train, validate and test GIGN. First, cd ./GIGN

1. Model training

Firstly, download the preprocessed datasets from https://zenodo.org/record/7490623#.Y60PTnZBxD8, and organize them as './data/train', './data/valid', './data/test2013/', './data/test2016/', and './data/test2019/'.
Secondly, run train.py using python train.py.

2. Model testing

Run test.py using python predict.py.
You may need to modify some file paths in the source code before running it.
You can also use python evaluate.py, which will return the mean (std) of performance for the three independent runs.

3. Process raw data

We provide a demo to explain how to process the raw data. This demo use ./data/toy_examples.csv and ./data/toy_set/ as examples.
Firstly, run preprocessing.py using python preprocessing.py.
Secondly, run dataset_GIGN.py using python dataset_GIGN.py.
Thirdly, run train.py using python train_example.py.

4. Test the trained model in other external test sets

Firstly, please organize the data as a structure similar to './data/toy_set' folder.
-data
  -external_test
      -pdb_id
        -pdb_id_ligand.mol2
        -pdb_id_protein.pdb
Secondly, run preprocessing.py using python preprocessing.py.
Thirdly, run dataset_GIGN.py using python dataset_GIGN.py.
Fourth, run predict.py using python predict.py.
You may need to modify some file paths in the source code before running it.

Other baselines:

The usage of other baselines is similar to GIGN.